DESRES-ANTON-10906555 2 µs simulations of 50 FDA approved or investigational drug molecules binding to a construct of the SARS-CoV-2 trimeric spike protein (2 µs )

D. E. Shaw Research

DESRES

50 2 µs trajectories of FDA approved or investigational drug molecules that in simulation remained bound to a construct of the SARS-CoV-2 trimeric spike protein at positions that might conceivably allosterically disrupt the interaction between these proteins. The small molecule drugs and their initial binding poses were chosen from a combination of molecular dynamics simulation and docking performed using an FDA-investigational drug library. The 50 putative spike protein binding small molecules located at three regions on the spike trimer, a pocket in the RBD whose formation may possibly enhance RBD-RBD interactions in the closed conformation (8 molecules), a pocket between the two RBDs in the closed conformation (29 molecules), and a pocket that involves three RBDs in the closed conformation (13 molecules). The simulations used the Amber ff99SB-ILDN force field for proteins, the CHARMM TIP3P model for water, and the generalized Amber force field for small molecules. The C- and N-peptide termini were capped with amide and acetyl groups respectively. The spike trimer construct was modeled from PDB entries 6VXX and 6VW1, only retaining the RBD and a short region from S1 fusion protein as a minimal system for maintaining a trimer assembly. The system was neutralized and salted with NaCl, with a final concentration of 0.15 M. The interval between frames is 1.2 ns. The simulations were conducted at 310 K in the NPT ensemble.

Input and Supporting Files:

DESRES-Trajectory_sarscov2-10906555-set_spike-structure.tar.gz

DESRES-Trajectory_sarscov2-10906555-set_spike-table.csv

DESRES-Trajectory_sarscov2-10906555.mp4

Trajectory: Get Trajectory (166 GB)

Represented Proteins: spike RBD

Represented Structures: 6vw1 6vxx

Models: SARS-CoV-2 trimeric spike protein binding to FDA approved or investigational drug molecules

DESRES-ANTON-10897850 10 µs simulation of of the trimeric SARS-CoV-2 spike glycoprotein in aqueous solution (10 µs )

D. E. Shaw Research

DESRES

A 10 µs simulation of the trimeric SARS-CoV-2 spike glycoprotein. System was initiated in a partially opened state (PDB entry 6VYB) which exhibited a high degree of conformational heterogeneity. In particular, the partially detached receptor binding domain sampled a variety of orientations, and further detached from the S2 fusion machinery. The simulation used the Amber ff99SB-ILDN force field for proteins, the CHARMM TIP3P model for water, and the generalized Amber force field for glycosylated asparagine. The C- and N-peptide termini, including those exposed due to missing loops in the published structural models, are capped with amide and acetyl groups respectively. The system was neutralized and salted with NaCl, with a final concentration of 0.15 M. The total number of atoms in the system was 715439 for the closed state. The interval between frames is 1.2 ns. The simulations were conducted at 310 K in the NPT ensemble. We have released new versions of these simulations with enhancements to the spike protein model in [DESRES-ANTON-11021566,11021571] (https://www.deshawresearch.com/downloads/download_trajectory_sarscov2.cgi/#DESRES-ANTON-11021566), since the one used in this simulation is incomplete in some of the disordered loop regions (i.e., resid 455 to 461, resid 469 to 488) and in glycan chains.

Input and Supporting Files:

DESRES-Trajectory_sarscov2-10897850-structure.tar.gz

DESRES-Trajectory_sarscov2-10897850.mp4

Trajectory: Get Trajectory (62 GB)

Represented Proteins: spike

Represented Structures: 6vyb

Models: Trimeric SARS-CoV-2 spike glycoprotein (open state) in aqueous solution

Folding@home simulations of the apo SARS-CoV-2 spike RBD (with glycosylation) (1.8 ms )

Ivy Zhang

Folding@home -- Chodera lab

All-atom MD simulations of the SARS-CoV-2 spike protein receptor binding domain (RBD) (with glycosylation), simulated using Folding@Home. Complete details of this simulation are available here. Brief details appear below. Publication: https://doi.org/10.1016/j.cell.2021.01.037 System preparation: The RBD complex was constructed from PDB ID 6M0J (Chain B). 6M0J was refined using ISOLDE to better fit the experimental electron density using detailed manual inspection. ACE2 (+ associated glycans) were then deleted. The equilibrated structure was then used to initiate parallel distributed MD simulations on Folding@home (Shirts and Pande, 2000, Zimmerman et al., 2020). Simulations were run with OpenMM 7.4.2 (Folding@home core22 0.0.13). Production simulations used the same Langevin integrator as the NPT equilibration described above. In total, 2995 independent MD simulations were generated on Folding@home. Conformational snapshots (frames) were stored at an interval of 1 ns/frame for subsequent analysis. The resulting final dataset contained 2995 trajectories, 1.8 ms of aggregate simulation time. Solute-only trajectories: The solute-only trajectories (with counterions) are available as MDTraj HDF5 files that contain both topology and trajectory information. A single trajectory (RUN0 CLONE0) can be downloaded using the AWS CLI:

aws s3 --no-sign-request cp s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-apo/munged/solute/17314/run0-clone0.h5 .

All HDF5 trajectories can be retrieved with

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-apo/munged/solute/17314 .

Entire dataset: The raw Folding@home dataset is made available through the AWS Open Data Registry and can be retrieved through the AWS CLI. The dataset consists of a single project (PROJ17314) and has a RUN*/CLONE*/result* directory structure. RUNs denote different equilibrated starting structures. CLONEs denote different independent replica trajectories. To retrieve raw trajectory files in gromacs XTC format for the whole dataset, you can use the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-apo/raw/PROJ17314 .

Folding@home initial files: System setup and input files can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-apo/setup-files/17314 .

Contributors: Ivy Zhang, William G. Glass, Tristan I. Croll, Aoife M. Harbison, Elisa Fadda, John D. Chodera. License: All data is freely available under the Creative Commons CC0 (“No Rights Reserved”) license.

Input and Supporting Files: ---

Trajectory: Get Trajectory (50 GB)

Represented Proteins: spike RBD

Represented Structures: 6m0j

Models: SARS-CoV-2 spike receptor-binding domain: ISOLDE refined model with N343 glycan

MMGB/SA Consensus Estimate of the Binding Free Energy Between the Novel Coronavirus Spike Protein to the Human ACE2 Receptor (50 ns )

Negin Forouzesh, Alexey Onufriev

California State University, Los Angeles and Virginia Tech

50 ns simulation trajectory of a truncated SARS-CoV-2 spike receptor binding domain the human ACE2 receptor. The simulations used the Amber ff14SB force field and the OPC water model. The initial structure (PDB ID:6m0j) was truncated in order to obtain a smaller complex feasible with the computational framework. A molecular mechanics generalized Born surface area (MMGB/SA) approach was employed to estimate absolute binding free energy of the truncated complex. The system was neutralized and salted with NaCl, with a final concentration of 0.15 M.The simulations were conducted at 300 K in the NPT ensemble.

Input and Supporting Files:

MD_Input

Trajectory: Get Trajectory (31 GB)

Represented Proteins: spike RBD ACE2

Represented Structures: 6m0j

Models: SARS-CoV-2 spike receptor-binding domain bound with ACE2

DESRES-ANTON-10897136 10 µs simulation of of the trimeric SARS-CoV-2 spike glycoprotein, no water or ions (10 µs )

D. E. Shaw Research

DESRES

A 10 µs simulation of the trimeric SARS-CoV-2 spike glycoprotein. System was initiated in the closed state (PDB entry 6VXX), which remained stable. The simulation used the Amber ff99SB-ILDN force field for proteins, the CHARMM TIP3P model for water, and the generalized Amber force field for glycosylated asparagine. The C- and N-peptide termini, including those exposed due to missing loops in the published structural models, are capped with amide and acetyl groups respectively. The system was neutralized and salted with NaCl, with a final concentration of 0.15 M. The total number of atoms in the system was 566502 for the closed state. The interval between frames is 1.2 ns. The simulations were conducted at 310 K in the NPT ensemble. We have released new versions of these simulations with enhancements to the spike protein model in [DESRES-ANTON-11021566,11021571] (https://www.deshawresearch.com/downloads/download_trajectory_sarscov2.cgi/#DESRES-ANTON-11021566), since the one used in this simulation is incomplete in some of the disordered loop regions (i.e., resid 455 to 461, resid 469 to 488) and in glycan chains.

Input and Supporting Files:

DESRES-Trajectory_sarscov2-10897136-structure.tar.gz

DESRES-Trajectory_sarscov2-10897136.mp4

Trajectory: Get Trajectory (4.1 GB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Trimeric SARS-CoV-2 spike glycoprotein (closed state) in aqueous solution

Riken CPR TMS, MD1_Up trajectory (1 microseconds )

Takaharu Mori, Jaewoon Jung, Chigusa Kobayashi, Hisham M. Dokainish, Suyong Re, Yuji Sugita

RIKEN CPR (Cluster for Pioneering Research), TMS (Theoretical molecular science) laboratory -- TMS (Theoretical molecular science) laboratory

The data set includes a trajectory file from Molecular Dynamics (MD) of a fully glycosylated SARS-CoV-2 S-protein in solution. Water molecules and counter ions were excluded. The starting structure is an active Up taken from CHARMM-GUI COVID-19 Archive (http://www.charmm-gui.org/docs/archive/covid19). We replaced counter ions K+ in the original model with Na+. The simulation used CHARMM36m force field for protein, and TIP3P water model. The simulation was performed using GENESIS. The coordinates were saved every 1 nanoseconds and aligned to S2 domain (Calpha atoms of residues 689-727, 854-1147).

Input and Supporting Files:

Up_pro-gly.psf

Trajectory: Get Trajectory (742MB)

Represented Proteins: spike

Represented Structures: 6vsb

Models:

Riken CPR TMS, MD2_Up trajectory (200 nanoseconds )

Takaharu Mori, Jaewoon Jung, Chigusa Kobayashi, Hisham M. Dokainish, Suyong Re, Yuji Sugita

RIKEN CPR (Cluster for Pioneering Research), TMS (Theoretical molecular science) laboratory -- TMS (Theoretical molecular science) laboratory

The data set includes a trajectory file from Molecular Dynamics (MD) of a fully glycosylated SARS-CoV-2 S-protein in solution. Water molecules and counter ions were excluded. The starting structure is an active Up taken from CHARMM-GUI COVID-19 Archive (http://www.charmm-gui.org/docs/archive/covid19). We replaced counter ions K+ in the original model with Na+. The simulation used CHARMM36m force field for protein, and TIP3P water model. The simulation was performed using GENESIS. The coordinates were saved every 1 nanoseconds and aligned to S2 domain (Calpha atoms of residues 689-727, 854-1147).

Input and Supporting Files:

Up_pro-gly.psf

Trajectory: Get Trajectory (149MB)

Represented Proteins: spike

Represented Structures: 6vsb

Models:

Clusters center of gREST from 1Up State simulations (300 ns )

sugita lab

CPR

PDB of cluster centers representing 13 clusters obtained from gREST_SSCR simulations starting from 1Up conformation. This includes clusters represent 1Up conformations(1Ua.pdb, 1Ub.pdb, 1Uc.pdb, 1Ue.pdb, 1Uf.pdb, 1Ug.pdb, 1Uh.pdb, 1Ui.pdb and 1Uj.pdb), clusters for 2Up like conformations (2Ula.pdb and 2Ulb.pdb)and 1Up/open conformation (1U_O.pdb).

Input and Supporting Files: ---

Trajectory: Get Trajectory (13.3 MB)

Represented Proteins: spike

Represented Structures: 6vyb

Models: Trimeric SARS-CoV-2 spike glycoprotein (1Up state) with and without simulation box

DESRES-ANTON-10897136 10 µs simulation of of the trimeric SARS-CoV-2 spike glycoprotein in aqueous solution (10 µs )

D. E. Shaw Research

DESRES

A 10 µs simulation of the trimeric SARS-CoV-2 spike glycoprotein. System was initiated in the closed state (PDB entry 6VXX), which remained stable. The simulation used the Amber ff99SB-ILDN force field for proteins, the CHARMM TIP3P model for water, and the generalized Amber force field for glycosylated asparagine. The C- and N-peptide termini, including those exposed due to missing loops in the published structural models, are capped with amide and acetyl groups respectively. The system was neutralized and salted with NaCl, with a final concentration of 0.15 M. The total number of atoms in the system was 566502 for the closed state. The interval between frames is 1.2 ns. The simulation was conducted at 310 K in the NPT ensemble. We have released new versions of these simulations with enhancements to the spike protein model in [DESRES-ANTON-11021566,11021571] (https://www.deshawresearch.com/downloads/download_trajectory_sarscov2.cgi/#DESRES-ANTON-11021566), since the one used in this simulation is incomplete in some of the disordered loop regions (i.e., resid 455 to 461, resid 469 to 488) and in glycan chains.

Input and Supporting Files:

DESRES-Trajectory_sarscov2-10897136-structure.tar.gz

DESRES-Trajectory_sarscov2-10897136.mp4

Trajectory: Get Trajectory (49 GB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Trimeric SARS-CoV-2 spike glycoprotein (closed state) in aqueous solution

SIRAH-CoV2 initiative - S1 Receptor Binding Domain in complex with human antibody CR3022 (12 µs )

Martin Soñora

Institut Pasteur de Montevideo -- Biomolecular Simulations Laboratory

This dataset contains the trajectory of a 12 microseconds-long coarse-grained molecular dynamics simulation of SARS-CoV-2 receptor binding domain in complex with a human antibody CR3022 (PDB id: 6W41). Simulations have been performed using the SIRAH force field running with the Amber18 package at the Uruguayan National Center for Supercomputing (ClusterUY) under the conditions reported in Machado et al. JCTC 2019, adding 150 mM NaCl according to Machado & Pantano JCTC 2020. Glycans have been removed from the structures.

The file 6W41_SIRAHcg_rawdata.tar contain all the raw information required to visualize (on VMD), analyze, backmap, and eventually continue the simulations using Amber18 or higher. Step-By-Step tutorials for running, visualizing, and analyzing CG trajectories using SirahTools can be found at SIRAH website.

Additionally, the file 6W41_SIRAHcg_12us_prot.tar contains only the protein coordinates, while 6W41_SIRAHcg_12us_prot_skip10ns.tar contains one frame every 10ns.

To take a quick look at the trajectory:

1- Untar the file 6W41_SIRAHcg_12us_prot_skip10ns.tar

2- Open the trajectory on VMD using the command line: vmd 6W41_SIRAHcg_prot.prmtop 6W41_SIRAHcg_prot_12us_skip10ns.ncrst 6W41_SIRAHcg_prot_12us_skip10ns.nc -e sirah_vmdtk.tcl

Note that you can use normal VMD drawing methods as vdw, licorice, etc., and coloring by restype, element, name, etc.

Input and Supporting Files: ---

Trajectory: Get Trajectory (20.6 GB)

Represented Proteins: spike RBD

Represented Structures: 6W41

Models: ---

Cluster ensemble of 1UP second populated cluster (300 ns )

sugita lab

CPR

30 PDB structures of the second populated cluster obtained from gREST_SSCR simulations starting from 1Up conformation. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (30.8 MB)

Represented Proteins: spike

Represented Structures: 6vyb

Models: Trimeric SARS-CoV-2 spike glycoprotein (1Up state) with and without simulation box

Interaction between the SARS-CoV-2 spike and the α7 nicotinic receptor (3 replicates, 300 ns each )

A.S.F. Oliveira

University of Bristol -- Mulholland Lab

MD simulation of the complex between the Y674-R685 region of the SARS-CoV-2 spike and the extracellular domain of the human α7 nicotinic acetylcholine receptor. A7_nAChR-spike.tar.gz contains all the following files. A7_nAChR-spike_complex.pdb A7_nAChR-spike_r1.tpr A7_nAChR-spike_r1.xtc A7_nAChR-spike_r2.tpr A7_nAChR-spike_r2.xtc A7_nAChR-spike_r3.tpr A7_nAChR-spike_r3.xtc

Input and Supporting Files:

A7_nAChR-spike.tar.gz

Trajectory: Get Trajectory (9 GB)

Represented Proteins: spike

Represented Structures: https://molssi-bioexcel-covid-19-structure-therapeutics-hub.s3.amazonaws.com/MulhollandGroup/nAChR-spike_interaction/A7_nAChR-spike_complex.pdb

Models: ---

Simulations of SARS-CoV and SARS-CoV-2 RBD with ACE2 possessing different patterns of glycosylation (2 µs )

Gumbart lab

Two-microsecond trajectories of the receptor-binding domains from SARS-CoV and SARS-CoV-2 spike protein bound to the human receptor, ACE2 with two distinct glycosylation schemes (three replicas each, joined in a single DCD file for each scheme) and with no glycans (two replicas each). Simulation systems were constructed with VMD, equilibrated initially with NAMD, and then run for 2 µs each with Amber16. Simulations used a 4-fs timestep enabled by hydrogen-mass repartitioning (HMR).

Input and Supporting Files: ---

Trajectory: Get Trajectory (22 GB)

Represented Proteins: spike RBD ACE2

Represented Structures: 2AJF 6M17

Models: ---

Riken CPR TMS, TMD1_toDown trajectory (20 nanoseconds )

Takaharu Mori, Jaewoon Jung, Chigusa Kobayashi, Hisham M. Dokainish, Suyong Re, Yuji Sugita

RIKEN CPR (Cluster for Pioneering Research), TMS (Theoretical molecular science) laboratory -- TMS (Theoretical molecular science) laboratory

The data set includes a trajectory file from Targeted Molecular Dynamics (TMD) simulations of a fully glycosylated SARS-CoV-2 S-protein in solution. Water molecules and counter ions were excluded. The data includes trajectory from TMD simulation of Up to Down forms. The simulations used CHARMM36m force field for protein, and TIP3P water model. The simulations were performed using GENESIS. The coordinates were saved every 1 nanoseconds and aligned to S2 domain (Calpha atoms of residues 689-727, 854-1147).

Input and Supporting Files:

Up_pro-gly.psf

Trajectory: Get Trajectory (15MB)

Represented Proteins: spike

Represented Structures: 6vsb 6vxx

Models:

Riken CPR TMS, TMD3_toDown trajectory (50 nanoseconds )

Takaharu Mori, Jaewoon Jung, Chigusa Kobayashi, Hisham M. Dokainish, Suyong Re, Yuji Sugita

RIKEN CPR (Cluster for Pioneering Research), TMS (Theoretical molecular science) laboratory -- TMS (Theoretical molecular science) laboratory

The data set includes a trajectory file from Targeted Molecular Dynamics (TMD) simulations of a fully glycosylated SARS-CoV-2 S-protein in solution. Water molecules and counter ions were excluded. The data includes trajectory from TMD simulation of Up to Down forms. The simulations used CHARMM36m force field for protein, and TIP3P water model. The simulations were performed using GENESIS. The coordinates were saved every 1 nanoseconds and aligned to S2 domain (Calpha atoms of residues 689-727, 854-1147).

Input and Supporting Files:

Up_pro-gly.psf

Trajectory: Get Trajectory (38MB)

Represented Proteins: spike

Represented Structures: 6vsb 6vxx

Models:

Riken CPR TMS, TMD3_toUp trajectory (50 nanoseconds )

Takaharu Mori, Jaewoon Jung, Chigusa Kobayashi, Hisham M. Dokainish, Suyong Re, Yuji Sugita

RIKEN CPR (Cluster for Pioneering Research), TMS (Theoretical molecular science) laboratory -- TMS (Theoretical molecular science) laboratory

The data set includes a trajectory file from Targeted Molecular Dynamics (TMD) simulations of a fully glycosylated SARS-CoV-2 S-protein in solution. Water molecules and counter ions were excluded. The data includes trajectory from TMD simulation of Down to Up forms. The simulations used CHARMM36m force field for protein, and TIP3P water model. The simulations were performed using GENESIS. The coordinates were saved every 1 nanoseconds and aligned to S2 domain (Calpha atoms of residues 689-727, 854-1147).

Input and Supporting Files:

Down_pro-gly.psf

Trajectory: Get Trajectory (38MB)

Represented Proteins: spike

Represented Structures: 6vxx 6vsb

Models:

Gromacs 60 ns MD of SARS-CoV-2 spike trimer, All Atom model (60 ns )

Dmitry Morozov

University of Jyvaskyla

This trajectory is from a 60 ns MD simulation of the SARS-CoV-2 spike protein. The protein was solvated in a 20 x 20 x 20 nm water box containing 0.1 M NaCl. The simulation was performed with Gromacs 2018.8 on the Puhti cluster located at the CSC-IT using the Charmm27 force field. The interval between frames is 80 ps. The simulation was conducted in the NPT ensemble (1 bar). This trajectory is all atom.

Input and Supporting Files:

trimer

Trajectory: Get Trajectory (2.0 GB)

Represented Proteins: spike

Represented Structures: 6VXX

Models: SARS-CoV-2 spike protein trimer (closed state) model for MD simulations

Folding@home simulations of the SARS-CoV-2 spike RBD bound to human ACE2 (725.3 µs )

Ivy Zhang

Folding@home -- Chodera lab

All-atom MD simulations of the SARS-CoV-2 spike protein receptor binding domain (RBD) bound to human angiotensin converting enzyme-related carboypeptidase (ACE2), simulated using Folding@Home. The “wild-type” RBD and three mutants (N439K, K417V, and the double mutant N439K/K417V) were simulated.

Complete details of this simulation are available here. Brief details appear below.

Publication: https://doi.org/10.1016/j.cell.2021.01.037

System preparation: The RBD:ACE2 complex was constructed from individual RBD (PDB: 6m0j, Chain E) and ACE2 (PDB: 1r42, Chain A) monomers aligned to the full RBD:ACE2 structure (PDB: 6m0j. These structural models were further refined by Tristan Croll using ISOLDE (Croll, 2018) and deposited in the Coronavirus Structural Taskforce (CST) database (Croll et al., 2020) to produce refined 6m0j and refined 1r42 models. The resulting RBD and ACE2 monomers were then aligned in PyMOL 2.3.2 to the CST 6m0j structure to create an initial RBD:ACE2 complex.

Full glycosylation patterns for ACE2 and RBD glycans were determined from Shajahan et al. For the constructed RBD:ACE2 complex, these included sites: N53, N90, N103, N322, N432, N546, and N690 on ACE2 and N343 on the RBD. Base NAG residues of each glycan structure (FA2, FA26G1, FA2, FA2, FA2G2, A2, FA2, FA2G2, respectively) were acquired from Elisa Fadda. Each glycan was then aligned to the corresponding NAG stub in the RBD:ACE2 model in and any resulting clashes were refined in ISOLDE. Full details of the glycosylation patterns / structures used and full workflow are available here.

Folding@home simulation: The equilibrated structure was then used to initiate parallel distributed MD simulations on Folding@home (Shirts and Pande, 2000, Zimmerman et al., 2020). Simulations were run with OpenMM 7.4.2 (Folding@home core22 0.0.13). Production simulations used the same Langevin integrator as the NpT equilibration described above. In total, 8000 independent MD simulations were generated on Folding@home. Conformational snapshots (frames) were stored at an interval of 0.5 ns/frame for subsequent analysis. The resulting final dataset contained 8000 trajectories, 725.3 us of aggregate simulation time, and 1450520 frames. Solute-only trajectories: The solute-only trajectories (with counterions) are available as MDTraj HDF5 files that contain both topology and trajectory information. A single trajectory of the WT RBD (RUN3) (~30 MB) can be downloaded using the AWS CLI:

aws s3 --no-sign-request cp s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-ACE2-RBD/munged/solute/PROJ17311/run3-clone0.h5 .

All HDF5 trajectories (~300 GB) can be retrieved with

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-ACE2-RBD/munged/solute/PROJ17311 .

Entire dataset: The raw Folding@home dataset is made available through the AWS Open Data Registry and can be retrieved through the AWS CLI. The dataset consists of a single project (PROJ17311) and has a RUN*/CLONE*/result* directory structure. RUNs denote different RBD mutants: N439K (RUN0), K417V (RUN1), N439K/K417V (RUN2), and WT (RUN3). CLONEs denote different independent replica trajectories.

To retrieve raw trajectory files in gromacs XTC format for the whole dataset (7 TB), you can use the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-ACE2-RBD/raw-data/PROJ17311 .

Folding@home initial files: System setup and input files can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-ACE2-RBD/setup/PROJ17311 .

Contributors: Ivy Zhang, William G. Glass, Tristan I. Croll, Aoife M. Harbison, Elisa Fadda, John D. Chodera.

License: All data is freely available under the Creative Commons CC0 (“No Rights Reserved”) license.

Input and Supporting Files: ---

Trajectory: Get Trajectory (341 GB)

Represented Proteins: spike RBD ACE2

Represented Structures: 6m0j 1R42

Models: SARS-CoV-2 spike receptor-binding domain: ISOLDE refined model with N343 glycan Native Human Angiotensin Converting Enzyme-Related Carboxypeptidase (ACE2): ISOLDE refined model with glycans

Riken CPR TMS, MD2_Down trajectory (200 nanoseconds )

Takaharu Mori, Jaewoon Jung, Chigusa Kobayashi, Hisham M. Dokainish, Suyong Re, Yuji Sugita

RIKEN CPR (Cluster for Pioneering Research), TMS (Theoretical molecular science) laboratory -- TMS (Theoretical molecular science) laboratory

The data set includes a trajectory file from Molecular Dynamics (MD) of a fully glycosylated SARS-CoV-2 S-protein in solution. Water molecules and counter ions were excluded. The starting structure is an inactive Down taken from CHARMM-GUI COVID-19 Archive (http://www.charmm-gui.org/docs/archive/covid19). We replaced counter ions K+ in the original model with Na+. The simulation used CHARMM36m force field for protein, and TIP3P water model. The simulation was performed using GENESIS. The coordinates were saved every 1 nanoseconds and aligned to S2 domain (Calpha atoms of residues 689-727, 854-1147).

Input and Supporting Files:

Down_pro-gly.psf

Trajectory: Get Trajectory (149MB)

Represented Proteins: spike

Represented Structures: 6vxx

Models:

Cluster ensemble of 1UP/open conformations (300 ns )

sugita lab

CPR

30 PDB structures of the 1Up/open conformations obtained from gREST_SSCR simulations starting from Up conformation. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (30.8 MB)

Represented Proteins: spike

Represented Structures: 6vyb

Models: Trimeric SARS-CoV-2 spike glycoprotein (1Up state) with and without simulation box

DESRES-ANTON-11021571 10 µs simulation of of the trimeric SARS-CoV-2 spike glycoprotein in aqueous solution (10 µs )

D. E. Shaw Research

DESRES

10 µs simulation trajectory of the trimeric SARS-CoV-2 spike glycoprotein with additional loop structures and glycan chains to improve the spike protein model originally released in DESRES-ANTON-[10897136,10897850]. Trajectory was initiated in a partially opened state (PDB entry 6VYB). The simulation used the Amber ff99SB-ILDN force field for proteins, the CHARMM TIP3P model for water, and the generalized Amber force field for glycosylated asparagine. The C- and N-peptide termini are capped with amide and acetyl groups respectively. The system was neutralized and salted with NaCl, with a final concentration of 0.15 M. The interval between frames is 1.2 ns. The simulations were conducted at 310 K in the NPT ensemble.

Input and Supporting Files:

DESRES-Trajectory_sarscov2-11021571-structure.tar.gz

DESRES-Trajectory_sarscov2-11021571.mp4

Trajectory: Get Trajectory (67 GB)

Represented Proteins: spike

Represented Structures: 6vyb

Models: Trimeric SARS-CoV-2 spike glycoprotein (open state) in aqueous solution

SIRAH-CoV2 initiative - Spike´s RBD/ACE2-B0AT1 complex (4 µs )

Florencia Klein

Institut Pasteur de Montevideo -- Biomolecular Simulations Laboratory

This dataset contains an trajectory of four microseconds-long coarse-grained molecular dynamics simulation of the hexameric complex between SARS-CoV2 Spike´s RBD, ACE2, and B0AT1 (PDB id: 6M17). Simulations have been performed using the SIRAH force field running with the Amber18 package at the Uruguayan National Center for Supercomputing (ClusterUY) under the conditions reported in Machado et al. JCTC 2019, adding 150 mM NaCl according to Machado & Pantano JCTC 2020. Zinc ions were parameterized as reported in Klein et al. 2020.

The files 6M17_SIRAHcg_rawdata_0-1.tar, 6M17_SIRAHcg_rawdata_1-2.tar, 6M17_SIRAHcg_rawdata_2-3.tar, and 6M17_SIRAHcg_rawdata_3-4.tar, contain all the raw information required to visualize (on VMD), analyze, backmap, and eventually continue the simulations using Amber18 or higher. Step-By-Step tutorials for running, visualizing, and analyzing CG trajectories using SirahTools can be found at SIRAH website.

Additionally, the file 6M17_SIRAHcg_4us_prot.tar contains only the protein coordinates, while 6M17_SIRAHcg_4us_prot_skip10ns.tar contains one frame every 10ns.

To take a quick look at the trajectory:

1- Untar the file 6M17_SIRAHcg_4us_prot_skip10ns.tar

2- Open the trajectory on VMD using the command line: vmd 6M17_SIRAHcg_prot.prmtop 6M17_SIRAHcg_prot_4us_skip10ns.ncrst 6M17_SIRAHcg_prot_4us_skip10ns.nc -e sirah_vmdtk.tcl

Note that you can use normal VMD drawing methods as vdw, licorice, etc., and coloring by restype, element, name, etc.

Input and Supporting Files: ---

Trajectory: Get Trajectory (20.1 GB)

Represented Proteins: spike RBD ACE2 BoAT1

Represented Structures: 6M17

Models: ---

SIRAH-CoV2 initiative - S2 Spike core fragment in postfusion state (10 µs )

Florencia Klein

Institut Pasteur de Montevideo -- Biomolecular Simulations Laboratory

This dataset contains the trajectory of a 10 microseconds-long coarse-grained molecular dynamics simulation of SARS-CoV2 Spike S2 fragment in its postfusion form (PDB id: 6M1V). Simulations have been performed using the SIRAH force field running with the Amber18 package at the Uruguayan National Center for Supercomputing (ClusterUY) under the conditions reported in Machado et al. JCTC 2019, adding 150 mM NaCl according to Machado & Pantano JCTC 2020.

The files 6M1V_SIRAHcg_rawdata_0-5us.tar, and 6M1V_SIRAHcg_rawdata_5-10us.tar contain all the raw information required to visualize (on VMD 1.9.3), analyze, backmap, and eventually continue the simulations using Amber18 or higher. Step-By-Step tutorials for running, visualizing, and analyzing CG trajectories using SirahTools can be found at SIRAH website.

Additionally, the file 6M1V_SIRAHcg_10us_prot.tar contains only the protein coordinates, while 6M1V_SIRAHcg_10us_prot_skip10ns.tar contains one frame every 10ns.

To take a quick look at the trajectory:

1- Untar the file 6M1V_SIRAHcg_10us_prot_skip10ns.tar

2- Open the trajectory on VMD using the command line: vmd 6M1V_SIRAHcg_prot.prmtop 6M1V_SIRAHcg_prot_10us_skip10ns.ncrst 6M1V_SIRAHcg_prot_10us_skip10ns.nc -e sirah_vmdtk.tcl

Note that you can use normal VMD drawing methods as vdw, licorice, etc., and coloring by restype, element, name, etc.

Input and Supporting Files: ---

Trajectory: Get Trajectory (17.8 GB)

Represented Proteins: spike S2

Represented Structures: 6M1V

Models: ---

Cluster ensemble of 1UP top populated cluster (300 ns )

sugita lab

CPR

30 PDB structures of the top populated cluster obtained from gREST_SSCR simulations starting from 1Up conformation. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (30.8 MB)

Represented Proteins: spike

Represented Structures: 6vyb

Models: Trimeric SARS-CoV-2 spike glycoprotein (1Up state) with and without simulation box

Clusters center of gREST from Down State simulations (500 ns )

sugita lab

CPR

PDB of cluster centers representing 13 clusters obtained from gREST_SSCR simulations starting from Down conformation. This includes Down symmetric (D1_Sym.pdb and D2_Sym.pdb), Down asymmetric (D1_asym.pdb and D2_asym.pdb), Intermediate 1 (I1a.pdb, I1b.pdb and I1c.pdb), Intermediate 2 (I2a.pdb, I2b.pdb and I2c.pdb), Intermediate 3 (I3a.pdb and I3b.pdb) and 1Up like (1U_L.pdb) conformations. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (13.3 MB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Trimeric SARS-CoV-2 spike glycoprotein (Down state) with and without simulation box

Cluster ensemble of Intermediate 2a (500 ns )

sugita lab

CPR

30 PDB structures of the intermediate (I2a) cluster obtained from gREST_SSCR simulations starting from Down conformation. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (30.8 MB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Trimeric SARS-CoV-2 spike glycoprotein (Down state) with and without simulation box

DESRES-ANTON-10906555 2 µs simulations of 50 FDA approved or investigational drug molecules binding to a construct of the SARS-CoV-2 trimeric spike protein, no water or ions (2 µs )

D. E. Shaw Research

DESRES

50 2 µs trajectories of FDA approved or investigational drug molecules that in simulation remained bound to a construct of the SARS-CoV-2 trimeric spike protein at positions that might conceivably allosterically disrupt the interaction between these proteins. The small molecule drugs and their initial binding poses were chosen from a combination of molecular dynamics simulation and docking performed using an FDA-investigational drug library. The 50 putative spike protein binding small molecules located at three regions on the spike trimer, a pocket in the RBD whose formation may possibly enhance RBD-RBD interactions in the closed conformation (8 molecules), a pocket between the two RBDs in the closed conformation (29 molecules), and a pocket that involves three RBDs in the closed conformation (13 molecules). The simulations used the Amber ff99SB-ILDN force field for proteins, the CHARMM TIP3P model for water, and the generalized Amber force field for small molecules. The C- and N-peptide termini were capped with amide and acetyl groups respectively. The spike trimer construct was modeled from PDB entries 6VXX and 6VW1, only retaining the RBD and a short region from S1 fusion protein as a minimal system for maintaining a trimer assembly. The system was neutralized and salted with NaCl, with a final concentration of 0.15 M. The interval between frames is 1.2 ns. The simulations were conducted at 310 K in the NPT ensemble.

Input and Supporting Files:

DESRES-Trajectory_sarscov2-10906555-set_spike-structure.tar.gz

DESRES-Trajectory_sarscov2-10906555-set_spike-table.csv

DESRES-Trajectory_sarscov2-10906555.mp4

Trajectory: Get Trajectory (14 GB)

Represented Proteins: spike RBD

Represented Structures: 6vw1 6vxx

Models: SARS-CoV-2 trimeric spike protein binding to FDA approved or investigational drug molecules

Folding@home simulations of the SARS-CoV-2 spike RBD with P337L mutation bound to monoclonal antibody S309 (923.2 µs )

Ivy Zhang

Folding@home -- Chodera lab

All-atom MD simulations of the SARS-CoV-2 spike protein receptor binding domain (RBD) with P337L mutation bound to monoclonal antibody S309, simulated using Folding@Home. Complete details of this simulation are available here. Brief details appear below. Publication: https://doi.org/10.1038/s41586-021-03807-6 System preparation: The RBD:S309 complex was constructed from PDB ID 7JX3 (Chains A, B, and R). 7JX3 was first refined using ISOLDE to better fit the experimental electron density using detailed manual inspection. Refinement included adjusting several rotamers, flipping several peptide bonds, fixing several weakly resolved waters, and building in a missing four-residue-long loop. Though the N343 glycan N-Acetylglucosamine (NAG) was present in 7JX3, ISOLDE was used to construct a complex glycan at N343. The full glycosylation pattern was determined from Shajahan et al. and Watanabe et al. The glycan structure used for N343 (FA2G2) corresponds to the most stable conformer obtained from multi microsecond molecular dynamics (MD) simulations of cumulative sampling. The base NAG residue in FA2G2 was aligned to the corresponding NAG stub in the RBD:S309 model and any resulting clashes were refined in ISOLDE. PyMOL was used to mutate RBD’s P337 to LEU. The equilibrated structure was then used to initiate parallel distributed MD simulations on Folding@home (Shirts and Pande, 2000, Zimmerman et al., 2020). Simulations were run with OpenMM 7.4.2 (Folding@home core22 0.0.13). Production simulations used the same Langevin integrator as the NPT equilibration described above. In total, 5985 independent MD simulations were generated on Folding@home. Conformational snapshots (frames) were stored at an interval of 1 ns/frame for subsequent analysis. The resulting final dataset contained 5985 trajectories, 923.2 µs of aggregate simulation time. Solute-only trajectories: The solute-only trajectories (with counterions) are available as MDTraj HDF5 files that contain both topology and trajectory information. A single trajectory (RUN0 CLONE0) can be downloaded using the AWS CLI:

aws s3 --no-sign-request cp s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/munged/solute/17343/run0-clone0.h5 .

All HDF5 trajectories can be retrieved with

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/munged/solute/17343 .

Entire dataset: The raw Folding@home dataset is made available through the AWS Open Data Registry and can be retrieved through the AWS CLI. The dataset consists of a single project (PROJ17343) and has a RUN*/CLONE*/result* directory structure. RUNs denote different equilibrated starting structures. CLONEs denote different independent replica trajectories. To retrieve raw trajectory files in gromacs XTC format for the whole dataset, you can use the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/raw-data/PROJ17343 .

Folding@home initial files: System setup and input files can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/setup-files/17343 .

Contributors: Ivy Zhang, William G. Glass, Tristan I. Croll, Aoife M. Harbison, Elisa Fadda, John D. Chodera. License: All data is freely available under the Creative Commons CC0 (“No Rights Reserved”) license.

Input and Supporting Files: ---

Trajectory: Get Trajectory (91 GB)

Represented Proteins: spike RBD

Represented Structures: 7jx3

Models: SARS-CoV-2 spike receptor-binding domain bound with S309: ISOLDE refined model with N343 glycan and P337L mutation

Riken CPR TMS, TMD1_toUp trajectory (20 nanoseconds )

Takaharu Mori, Jaewoon Jung, Chigusa Kobayashi, Hisham M. Dokainish, Suyong Re, Yuji Sugita

RIKEN CPR (Cluster for Pioneering Research), TMS (Theoretical molecular science) laboratory -- TMS (Theoretical molecular science) laboratory

The data set includes a trajectory file from Targeted Molecular Dynamics (TMD) simulations of a fully glycosylated SARS-CoV-2 S-protein in solution. Water molecules and counter ions were excluded. The data includes trajectory from TMD simulation of Down to Up forms. The simulations used CHARMM36m force field for protein, and TIP3P water model. The simulations were performed using GENESIS. The coordinates were saved every 1 nanoseconds and aligned to S2 domain (Calpha atoms of residues 689-727, 854-1147).

Input and Supporting Files:

Down_pro-gly.psf

Trajectory: Get Trajectory (15MB)

Represented Proteins: spike

Represented Structures: 6vxx 6vsb

Models:

Trajectory of the Spike protein in complex with human ACE2 (50 ns )

Oostenbrink Lab

University of Natural Resources and Life Sciences, Vienna

Atomistic MD simulations of the Spike protein in complex with the human ACE2 receptor, most probale glycosylations are added.

Input and Supporting Files:

inputdata.tar.gz

Trajectory: Get Trajectory (43 GB)

Represented Proteins: spike ACE2

Represented Structures: 6vyb 6m17

Models: Spike protein in complex with human ACE2

Cluster ensemble of Down asymmetric (500 ns )

sugita lab

CPR

30 PDB structures of the Down asymmetric (D1_asym) cluster obtained from gREST_SSCR simulations starting from Down conformation. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (30.8 MB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Trimeric SARS-CoV-2 spike glycoprotein (Down state) with and without simulation box

Cluster ensemble of Down symmetric (500 ns )

sugita lab

CPR

30 PDB structures of the Down symmetric (D1_Sym) cluster obtained from gREST_SSCR simulations starting from Down conformation. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (30.8 MB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Trimeric SARS-CoV-2 spike glycoprotein (Down state) with and without simulation box

1 microsecond trajecotry of glycosylated spike protein in open state for pdb:6VSB embedded in viral membrane (1 µs )

Klauda lab

All atom simulation of full-glycosylated spike protein in open state (pdb:6VSB) embedded in viral membrane. The structure was taken from Charmm-Gui at http://www.charmm-gui.org/?doc=archive&lib=covid19 where 8 models were built for the open state. For MD simulations we used model 1-2-1 provided by Im et. al. The PSF, PDB and XTC files are uploaded

Input and Supporting Files: ---

Trajectory: Get Trajectory (12 GB)

Represented Proteins: spike

Represented Structures: 6VSB

Models:

PMF calculations of SARS-CoV-2 spike opening

Gumbart lab

Conformations (~500) along the opening paths of the SARS-CoV-2 spike trimer with and without glycans as well as with the diproline mutation. Simulation systems were constructed with VMD, equilibrated initially with NAMD, and then used for two-dimensional replica-exchange umbrella sampling. Conformations provided here are taken from the minimum free-energy path between 1-RBD up and down states in each potential of mean force (PMF). Note that each DCD does not represent a continuous simulation trajecotry. Simulations used a 4-fs timestep enabled by hydrogen-mass repartitioning (HMR).

Input and Supporting Files: ---

Trajectory: Get Trajectory (962 MB)

Represented Proteins: spike RBD ACE2

Represented Structures: 6VYB 6XR8

Models: ---

Folding@home simulations of the SARS-CoV-2 spike RBD bound to monoclonal antibody S2H97 (623.7 us )

Ivy Zhang

Folding@home -- Chodera lab

All-atom MD simulations of the SARS-CoV-2 spike protein receptor binding domain (RBD) bound to monoclonal antibody S2H97, simulated using Folding@Home. Complete details of this simulation are available here. Brief details appear below. Publication: https://doi.org/10.1038/s41586-021-03807-6 System preparation: The RBD:S2H97 complex was constructed from PDB ID 7M7W (Chains S, C, and D). 7M7W was first refined using ISOLDE to better fit the experimental electron density using detailed manual inspection. Refinement included building in a missing four-residue-long loop. ISOLDE was used to construct a complex glycan at N343. The full glycosylation pattern was determined from Shajahan et al. and Watanabe et al. The glycan structure used for N343 (FA2G2) corresponds to the most stable conformer obtained from multi microsecond molecular dynamics (MD) simulations of cumulative sampling. The equilibrated structure was then used to initiate parallel distributed MD simulations on Folding@home (Shirts and Pande, 2000, Zimmerman et al., 2020). Simulations were run with OpenMM 7.4.2 (Folding@home core22 0.0.13). Production simulations used the same Langevin integrator as the NpT equilibration described above. In total, 4985 independent MD simulations were generated on Folding@home. Conformational snapshots (frames) were stored at an interval of 1 ns/frame for subsequent analysis. The resulting final dataset contained 4985 trajectories, 623.7 us of aggregate simulation time. Solute-only trajectories: The solute-only trajectories (with counterions) are available as MDTraj HDF5 files that contain both topology and trajectory information. A single trajectory (RUN0 CLONE0) (~29 MB) can be downloaded using the AWS CLI:

aws s3 --no-sign-request cp s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/munged/solute/17347/run0-clone0.h5 .

All HDF5 trajectories can be retrieved with

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/munged/solute/17347 .

Entire dataset: The raw Folding@home dataset is made available through the AWS Open Data Registry and can be retrieved through the AWS CLI. The dataset consists of a single project (PROJ17347) and has a RUN*/CLONE*/result* directory structure. RUNs denote different equilibrated starting structures. CLONEs denote different independent replica trajectories. To retrieve raw trajectory files in gromacs XTC format for the whole dataset, you can use the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/raw-data/PROJ17347 .

Folding@home initial files: System setup and input files can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/setup-files/17347 .

Contributors: Ivy Zhang, William G. Glass, Tristan I. Croll, Aoife M. Harbison, Elisa Fadda, John D. Chodera. License: All data is freely available under the Creative Commons CC0 (“No Rights Reserved”) license.

Input and Supporting Files: ---

Trajectory: Get Trajectory (60 GB)

Represented Proteins: spike RBD

Represented Structures: 7m7w

Models: SARS-CoV-2 spike receptor-binding domain bound with S2H97: ISOLDE refined model with N343 glycan

Trajectories of full-length SPIKE protein in the Open state (N165A / N234A mutations). (4.2 µs )

Amaro Lab

All-atom MD simulations of full-length SPIKE protein in the Open state bearing N165A and N234A mutations, protein + glycans only (not aligned). PSF and DCDs files are provided.

Input and Supporting Files: ---

Trajectory: Get Trajectory (31 GB)

Represented Proteins: spike

Represented Structures: 6VSB

Models:

Nonequilibrium simulations of the SARS-Cov-2 wild-type and D614G spike (180 replicates, 5 ns each )

A.S.F. Oliveira

University of Bristol -- Mulholland Lab

Nonequilibrium MD simulation of the unglycosylated and uncleaved ectodomain of the SARS-CoV-2 wild-type and D614G spike

Input and Supporting Files:

nonequilibrium_simulations.tar.gz

Trajectory: Get Trajectory (23 GB)

Represented Proteins: spike

Represented Structures: https://www.rcsb.org/structure/6ZB5

Models: ---

Trajectories of full-length SPIKE protein in the Open state. (4.2 µs )

Amaro Lab

All-atom MD simulations of full-length SPIKE protein in the Open state, protein + glycans only (not aligned). PSF and DCDs files are provided.

Input and Supporting Files: ---

Trajectory: Get Trajectory (31 GB)

Represented Proteins: spike

Represented Structures: 6VSB

Models:

DESRES-ANTON-11021566 10 µs simulation of of the trimeric SARS-CoV-2 spike glycoprotein in aqueous solution (10 µs )

D. E. Shaw Research

DESRES

10 µs simulation trajectory of the trimeric SARS-CoV-2 spike glycoprotein with additional loop structures and glycan chains to improve the spike protein model originally released in DESRES-ANTON-[10897136,10897850]. Trajectory was initiated in the closed state (PDB entry 6VXX). The simulation used the Amber ff99SB-ILDN force field for proteins, the CHARMM TIP3P model for water, and the generalized Amber force field for glycosylated asparagine. The C- and N-peptide termini are capped with amide and acetyl groups respectively. The system was neutralized and salted with NaCl, with a final concentration of 0.15 M. The interval between frames is 1.2 ns. The simulations were conducted at 310 K in the NPT ensemble.

Input and Supporting Files:

DESRES-Trajectory_sarscov2-11021566-structure.tar.gz

DESRES-Trajectory_sarscov2-11021566.mp4

Trajectory: Get Trajectory (51 GB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Improved trimeric SARS-CoV-2 spike glycoprotein (closed state) in aqueous solution

Folding@home simulations of the apo SARS-CoV-2 spike RBD (without glycosylation) (1.9 ms )

Ivy Zhang

Folding@home -- Chodera lab

All-atom MD simulations of the SARS-CoV-2 spike protein receptor binding domain (RBD) (without glycosylation), simulated using Folding@Home. Complete details of this simulation are available here. Brief details appear below. Publication: https://doi.org/10.1016/j.cell.2021.01.037 System preparation: The RBD complex was constructed from PDB ID 6M0J (Chain B). 6M0J was refined using ISOLDE to better fit the experimental electron density using detailed manual inspection. The N343 glycan and ACE2 (+ associated glycans) were then deleted. The equilibrated structure was then used to initiate parallel distributed MD simulations on Folding@home (Shirts and Pande, 2000, Zimmerman et al., 2020). Simulations were run with OpenMM 7.4.2 (Folding@home core22 0.0.13). Production simulations used the same Langevin integrator as the NPT equilibration described above. In total, 2995 independent MD simulations were generated on Folding@home. Conformational snapshots (frames) were stored at an interval of 1 ns/frame for subsequent analysis. The resulting final dataset contained 2995 trajectories, 1.9 ms of aggregate simulation time. Solute-only trajectories: The solute-only trajectories (with counterions) are available as MDTraj HDF5 files that contain both topology and trajectory information. A single trajectory (RUN0 CLONE0) can be downloaded using the AWS CLI:

aws s3 --no-sign-request cp s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-apo/munged/solute/17313/run0-clone0.h5 .

All HDF5 trajectories can be retrieved with

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-apo/munged/solute/17313 .

Entire dataset: The raw Folding@home dataset is made available through the AWS Open Data Registry and can be retrieved through the AWS CLI. The dataset consists of a single project (PROJ17313) and has a RUN*/CLONE*/result* directory structure. RUNs denote different equilibrated starting structures. CLONEs denote different independent replica trajectories. To retrieve raw trajectory files in gromacs XTC format for the whole dataset, you can use the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-apo/raw/PROJ17313 .

Folding@home initial files: System setup and input files can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-apo/setup-files/17313 .

Contributors: Ivy Zhang, William G. Glass, Tristan I. Croll, Aoife M. Harbison, Elisa Fadda, John D. Chodera. License: All data is freely available under the Creative Commons CC0 (“No Rights Reserved”) license.

Input and Supporting Files: ---

Trajectory: Get Trajectory (49 GB)

Represented Proteins: spike RBD

Represented Structures: 6m0j

Models: SARS-CoV-2 spike receptor-binding domain: ISOLDE refined model without N343 glycan

Interaction between the SARS-CoV-2 spike and the α4β2 nicotinic receptor (3 replicates, 300 ns each )

A.S.F. Oliveira

University of Bristol -- Mulholland Lab

MD simulation of the complex between the Y674-R685 region of the SARS-CoV-2 spike and the extracellular domain of the human α4β2 nicotinic acetylcholine receptor. A4B2_nAChR-spike.tar.gz contains the following files. A4B2_nAChR-spike_complex.pdb A4B2_nAChR-spike_r1.tpr A4B2_nAChR-spike_r1.xtc A4B2_nAChR-spike_r2.tpr A4B2_nAChR-spike_r2.xtc A4B2_nAChR-spike_r3.tpr A4B2_nAChR-spike_r3.xtc

Input and Supporting Files:

A4B2_nAChR-spike.tar.gz

Trajectory: Get Trajectory (9 GB)

Represented Proteins: spike

Represented Structures: https://molssi-bioexcel-covid-19-structure-therapeutics-hub.s3.amazonaws.com/MulhollandGroup/nAChR-spike_interaction/A4B2_nAChR-spike_complex.pdb

Models: ---

DESRES-ANTON-11021566 10 µs simulation of of the trimeric SARS-CoV-2 spike glycoprotein, no water or ions (10 µs )

D. E. Shaw Research

DESRES

10 µs simulation trajectory of the trimeric SARS-CoV-2 spike glycoprotein with additional loop structures and glycan chains to improve the spike protein model originally released in DESRES-ANTON-[10897136,10897850]. Trajectory was initiated in the closed state (PDB entry 6VXX). The simulation used the Amber ff99SB-ILDN force field for proteins, the CHARMM TIP3P model for water, and the generalized Amber force field for glycosylated asparagine. The C- and N-peptide termini are capped with amide and acetyl groups respectively. The system was neutralized and salted with NaCl, with a final concentration of 0.15 M. The interval between frames is 1.2 ns. The simulations were conducted at 310 K in the NPT ensemble.

Input and Supporting Files:

DESRES-Trajectory_sarscov2-11021566-structure.tar.gz

DESRES-Trajectory_sarscov2-11021566.mp4

Trajectory: Get Trajectory (5.3 GB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Improved trimeric SARS-CoV-2 spike glycoprotein (closed state) in aqueous solution

SIRAH-CoV2 initiative - RBD triple glycosylated at Asn331, 343, and 481 (10 µs )

Garay Pablo

Institut Pasteur de Montevideo -- Biomolecular Simulations Laboratory

This dataset contains the trajectory of a 10 microseconds-long coarse-grained molecular dynamics simulation of a Spike’s RBD from SARS-CoV2 glycosylated at Asn331, 343, and 481 with Man9 glycosylation trees. The initial coordinates correspond to amino acids 327 to 532 taken from the PDB structure 6XEY. Missing loops and glycosylation trees were added with CHARMM-GUI. Simulations have been performed using the SIRAH force field running with the Amber18 package at the Uruguayan National Center for Supercomputing (ClusterUY) under the conditions reported in Machado et al. JCTC 2019, adding 150 mM NaCl according to Machado & Pantano JCTC 2020. Glycan were parameterized as reported in Garay et at. 2020.

The files 6XEY-RBD-3Man9_SIRAHcg_0-4us.tar, 6XEY-RBD-3Man9_SIRAHcg_4-8us.tar, and 6XEY-RBD-3Man9_SIRAHcg_8-10us.tar, contain all the raw information required to visualize (on VMD), analyze, backmap the simulations. Step-By-Step tutorials for running, visualizing, and analyzing CG trajectories using SirahTools can be found at SIRAH website.

Additionally, the file with names ending in 6XEY-RBD-3Man9_SIRAHcg_glycoprot_10us.tar contains only the protein coordinates, while 6XEY-RBD-3Man9_SIRAHcg_glycoprot_skip10ns.tar contains one frame every 10ns.

To take a quick look at a the trajectory:

1- Untar the file 6XEY-RBD-3Man9_SIRAHcg_glycoprot_skip10ns.tar

2- Open the trajectory on VMD using the command line: vmd 6XEY-RBD-3Man9_SIRAHcg_glycoprot.prmtop 6XEY-RBD-3Man9_SIRAHcg_10us_skip10ns.ncrst 6XEY-RBD-3Man9_SIRAHcg_10us_skip10ns.nc -e sirah_vmdtk.tcl

Note that you can use normal VMD drawing methods as vdw, licorice, etc., and coloring by restype, element, name, etc.

Input and Supporting Files: ---

Trajectory: Get Trajectory (11.2 GB)

Represented Proteins: spike RBD

Represented Structures: 6XEY

Models: ---

1 microsecond trajecotry of glycosylated spike protein in closed state for pdb:6VXX embedded in viral membrane (1 µs )

Klauda lab

All atom simulation of full-glycosylated spike protein in closed state (pdb:6VXX) embedded in viral membrane. The structure was taken from Charmm-Gui at http://www.charmm-gui.org/?doc=archive&lib=covid19 where 8 models were built for the closed state. For MD simulations we used model 1-2-1 provided by Im et. al. The PSF, PDB and XTC files are uploaded

Input and Supporting Files: ---

Trajectory: Get Trajectory (12 GB)

Represented Proteins: spike

Represented Structures: 6VXX

Models:

Interaction between the SARS-CoV-2 spike and the αβγδ nicotinic receptor (3 replicates, 300 ns each )

A.S.F. Oliveira

University of Bristol -- Mulholland Lab

MD simulation of the complex between the Y674-R685 region of the SARS-CoV-2 spike and the extracellular domain of the αβγδ nicotinic acetylcholine receptor from Tetronarce californica (formerly Torpedo californica). ABGD_nAChR-spike.tar.gz contains the following files ABGD_nAChR-spike_complex.pdb ABGD_nAChR-spike_r1.tpr ABGD_nAChR-spike_r1.xtc ABGD_nAChR-spike_r2.tpr ABGD_nAChR-spike_r2.xtc ABGD_nAChR-spike_r3.tpr ABGD_nAChR-spike_r3.xtc

Input and Supporting Files:

ABGD_nAChR-spike.tar.gz

Trajectory: Get Trajectory (9 GB)

Represented Proteins: spike

Represented Structures: https://molssi-bioexcel-covid-19-structure-therapeutics-hub.s3.amazonaws.com/MulhollandGroup/nAChR-spike_interaction/ABGD_nAChR-spike_complex.pdb

Models: ---

Folding@home simulations of the SARS-CoV-2 spike protein (1.2 ms )

Maxwell Zimmerman

Folding@home -- Bowman lab

All-atom MD simulations of the SARS-CoV-2 spike protein, simulated using Folding@Home. The dataset comprises 3 projects, each having a RUN*/CLONE*/result* directory structure. Simulations were run using GROMACS (PROJ14217) or OpenMM (PROJ14235 and PROJ14561) and are stored as compressed binary XTC files. Each RUN represents a unique starting conformation, each CLONE is a unique MD run from the specified starting conformation, and each result is a fragment of the contiguous simulation. PROJ14217 and PROJ14253 were seeded using FAST simulations.

Topology files: The topology used in the trajectories can be downloaded directly here: PDB.

Entire dataset: The dataset is made available through the AWS Open Data Registry and can be retrieved through the AWS CLI. To retrieve raw trajectory files in gromacs XTC format for the whole dataset (7 TB), you can use the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-cryptic-pockets/SARS-CoV-2/spike/PROJ14217 .
aws s3 --no-sign-request sync s3://fah-public-data-covid19-cryptic-pockets/SARS-CoV-2/spike/PROJ14253 .
aws s3 --no-sign-request sync s3://fah-public-data-covid19-cryptic-pockets/SARS-CoV-2/spike/PROJ14561 .

Markov State Model: A polished Markov State Model (MSM), including representative cluster centers, transition probabilities, and equilibrum populations, can be downloaded using the AWS CLI. Details of how the MSM model was constructed can be found here.

aws s3 --no-sign-request sync s3://fah-public-data-covid19-cryptic-pockets/SARS-CoV-2/final_models/spike/model .

MSM cluster centers can be obtained as a gromacs XTC file from this URL: cluster centers XTC

Input files: System setup and input files can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-cryptic-pockets/SARS-CoV-2/spike/input_files .
aws s3 --no-sign-request sync s3://fah-public-data-covid19-cryptic-pockets/SARS-CoV-2/spike/PROJ14217_tpr_files .

FAST simulations: FAST simulations, which were used as seeds for Folding@Home simulations, can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-cryptic-pockets/SARS-CoV-2/FAST_simulations .

Input and Supporting Files: ---

Trajectory: Get Trajectory (6.5 TB)

Represented Proteins: spike

Represented Structures: 6VXX

Models: ---

Riken CPR TMS, TMD2_toUp trajectory (20 nanoseconds )

Takaharu Mori, Jaewoon Jung, Chigusa Kobayashi, Hisham M. Dokainish, Suyong Re, Yuji Sugita

RIKEN CPR (Cluster for Pioneering Research), TMS (Theoretical molecular science) laboratory -- TMS (Theoretical molecular science) laboratory

The data set includes a trajectory file from Targeted Molecular Dynamics (TMD) simulations of a fully glycosylated SARS-CoV-2 S-protein in solution. Water molecules and counter ions were excluded. The data includes trajectory from TMD simulation of Down to Up forms. The simulations used CHARMM36m force field for protein, and TIP3P water model. The simulations were performed using GENESIS. The coordinates were saved every 1 nanoseconds and aligned to S2 domain (Calpha atoms of residues 689-727, 854-1147).

Input and Supporting Files:

Down_pro-gly.psf

Trajectory: Get Trajectory (15MB)

Represented Proteins: spike

Represented Structures: 6vxx 6vsb

Models:

DESRES-ANTON-11021571 10 µs simulation of of the trimeric SARS-CoV-2 spike glycoprotein, no water or ions (10 µs )

D. E. Shaw Research

DESRES

10 µs simulation trajectory of the trimeric SARS-CoV-2 spike glycoprotein with additional loop structures and glycan chains to improve the spike protein model originally released in DESRES-ANTON-[10897136,10897850]. Trajectory was initiated in a partially opened state (PDB entry 6VYB). The simulation used the Amber ff99SB-ILDN force field for proteins, the CHARMM TIP3P model for water, and the generalized Amber force field for glycosylated asparagine. The C- and N-peptide termini are capped with amide and acetyl groups respectively. The system was neutralized and salted with NaCl, with a final concentration of 0.15 M. The interval between frames is 1.2 ns. The simulations were conducted at 310 K in the NPT ensemble.

Input and Supporting Files:

DESRES-Trajectory_sarscov2-11021571-structure.tar.gz

DESRES-Trajectory_sarscov2-11021571.mp4

Trajectory: Get Trajectory (5.3 GB)

Represented Proteins: spike

Represented Structures: 6vyb

Models: Trimeric SARS-CoV-2 spike glycoprotein (open state) in aqueous solution

SIRAH-CoV2 initiative - Glycosylated RBD (10 µs )

Garay Pablo

Institut Pasteur de Montevideo -- Biomolecular Simulations Laboratory

This dataset contains the trajectories of 10 microseconds-long coarse-grained molecular dynamics simulations of SARS-CoV2 Spike´s RBD glycosylated at Asn331 and Asn343. The initial coordinates correspond to amino acids 327 to 532 taken from the PDB structure 6VSB. Missing loops and glycosylation trees were added with CHARMM-GUI.

There are two different sets of simulations corresponding to Core Complex and High Mannose. Simulations have been performed using the SIRAH force field running with the Amber18 package at the Uruguayan National Center for Supercomputing (ClusterUY) under the conditions reported in Machado et al. JCTC 2019, adding 150 mM NaCl according to Machado & Pantano JCTC 2020. Glycan were parameterized as reported in Garay et at. 2020.

The files RBD-Man9_SIRAHcg_rawdata_0-6us.tar and RBD-Man9_SIRAHcg_rawdata_6-10us.tar, contain all the raw information required to visualize (on VMD), analyze, backmap the simulations. Analogous information for Core-complex glycosylations is contained in files RBD-Core-complex_SIRAHcg_rawdata_0-6us.tar and RBD-Core-complex_SIRAHcg_rawdata_6-10us.tar.

Step-By-Step tutorials for running, visualizing, and analyzing CG trajectories using SirahTools can be found at SIRAH website.

Additionally, the file with names ending in SIRAHcg_10us_prot.tar contains only the protein coordinates, while SIRAHcg_10us_prot_skip10ns.tar contains one frame every 10ns.

To take a quick look at a the trajectory:

1- Untar the file RBD-Core-complex_SIRAHcg_10us_prot_skip10ns.tar

2- Open the trajectory on VMD using the command line: vmd RBD-Core-complex_SIRAHcg_prot.prmtop RBD-Core-complex_SIRAHcg_prot_10us_skip10ns.ncrst RBD-Core-complex_SIRAHcg_prot_10us_skip10ns.nc -e sirah_vmdtk.tcl

Note that you can use normal VMD drawing methods as vdw, licorice, etc., and coloring by restype, element, name, etc.

Input and Supporting Files: ---

Trajectory: Get Trajectory (16.4 GB)

Represented Proteins: spike RBD

Represented Structures: 6VSB

Models: ---

Cluster ensemble of 2UP like conformations (300 ns )

sugita lab

CPR

30 PDB structures of the 2Up like conformations obtained from gREST_SSCR simulations starting from 1Up conformation. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (30.8 MB)

Represented Proteins: spike

Represented Structures: 6vyb

Models: Trimeric SARS-CoV-2 spike glycoprotein (1Up state) with and without simulation box

Folding@home simulations of the SARS-CoV-2 spike RBD with P337A mutation bound to monoclonal antibody S309 (907.0 µs )

Ivy Zhang

Folding@home -- Chodera lab

All-atom MD simulations of the SARS-CoV-2 spike protein receptor binding domain (RBD) with P337A mutation bound to monoclonal antibody S309, simulated using Folding@Home. Complete details of this simulation are available here. Brief details appear below. Publication: https://doi.org/10.1038/s41586-021-03807-6 System preparation: The RBD:S309 complex was constructed from PDB ID 7JX3 (Chains A, B, and R). 7JX3 was first refined using ISOLDE to better fit the experimental electron density using detailed manual inspection. Refinement included adjusting several rotamers, flipping several peptide bonds, fixing several weakly resolved waters, and building in a missing four-residue-long loop. Though the N343 glycan N-Acetylglucosamine (NAG) was present in 7JX3, ISOLDE was used to construct a complex glycan at N343. The full glycosylation pattern was determined from Shajahan et al. and Watanabe et al. The glycan structure used for N343 (FA2G2) corresponds to the most stable conformer obtained from multi microsecond molecular dynamics (MD) simulations of cumulative sampling. The base NAG residue in FA2G2 was aligned to the corresponding NAG stub in the RBD:S309 model and any resulting clashes were refined in ISOLDE. PyMOL was used to mutate RBD’s P337 to ALA. The equilibrated structure was then used to initiate parallel distributed MD simulations on Folding@home (Shirts and Pande, 2000, Zimmerman et al., 2020). Simulations were run with OpenMM 7.4.2 (Folding@home core22 0.0.13). Production simulations used the same Langevin integrator as the NPT equilibration described above. In total, 4998 independent MD simulations were generated on Folding@home. Conformational snapshots (frames) were stored at an interval of 1 ns/frame for subsequent analysis. The resulting final dataset contained 4998 trajectories, 907.0 µs of aggregate simulation time. Solute-only trajectories: The solute-only trajectories (with counterions) are available as MDTraj HDF5 files that contain both topology and trajectory information. A single trajectory (RUN0 CLONE0) can be downloaded using the AWS CLI:

aws s3 --no-sign-request cp s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/munged/solute/17342/run0-clone0.h5 .

All HDF5 trajectories can be retrieved with

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/munged/solute/17342 .

Entire dataset: The raw Folding@home dataset is made available through the AWS Open Data Registry and can be retrieved through the AWS CLI. The dataset consists of a single project (PROJ17342) and has a RUN*/CLONE*/result* directory structure. RUNs denote different equilibrated starting structures. CLONEs denote different independent replica trajectories. To retrieve raw trajectory files in gromacs XTC format for the whole dataset, you can use the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/raw-data/PROJ17342 .

Folding@home initial files: System setup and input files can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/setup-files/17342 .

Contributors: Ivy Zhang, William G. Glass, Tristan I. Croll, Aoife M. Harbison, Elisa Fadda, John D. Chodera. License: All data is freely available under the Creative Commons CC0 (“No Rights Reserved”) license.

Input and Supporting Files: ---

Trajectory: Get Trajectory (89 GB)

Represented Proteins: spike RBD

Represented Structures: 7jx3

Models: SARS-CoV-2 spike receptor-binding domain bound with S309: ISOLDE refined model with N343 glycan and P337A mutation

Cluster ensemble of 1UP like conformation (500 ns )

sugita lab

CPR

30 PDB structures of the 1Up like cluster obtained from gREST_SSCR simulations starting from Down conformation. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (30.8 MB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Trimeric SARS-CoV-2 spike glycoprotein (Down state) with and without simulation box

Folding@home simulations of the SARS-CoV-2 spike RBD bound to monoclonal antibody S309 (1.1 ms )

Ivy Zhang

Folding@home -- Chodera lab

All-atom MD simulations of the SARS-CoV-2 spike protein receptor binding domain (RBD) bound to monoclonal antibody S309, simulated using Folding@Home. Complete details of this simulation are available here. Brief details appear below. Publication: https://doi.org/10.1038/s41586-021-03807-6 System preparation: The RBD:S309 complex was constructed from PDB ID 7JX3 (Chains A, B, and R). 7JX3 was first refined using ISOLDE to better fit the experimental electron density using detailed manual inspection. Refinement included adjusting several rotamers, flipping several peptide bonds, fixing several weakly resolved waters, and building in a missing four-residue-long loop. Though the N343 glycan N-Acetylglucosamine (NAG) was present in 7JX3, ISOLDE was used to construct a complex glycan at N343. The full glycosylation pattern was determined from Shajahan et al. and Watanabe et al. The glycan structure used for N343 (FA2G2) corresponds to the most stable conformer obtained from multi microsecond molecular dynamics (MD) simulations of cumulative sampling. The base NAG residue in FA2G2 was aligned to the corresponding NAG stub in the RBD:S309 model and any resulting clashes were refined in ISOLDE. The equilibrated structure was then used to initiate parallel distributed MD simulations on Folding@home (Shirts and Pande, 2000, Zimmerman et al., 2020). Simulations were run with OpenMM 7.4.2 (Folding@home core22 0.0.13). Production simulations used the same Langevin integrator as the NpT equilibration described above. In total, 5000 independent MD simulations were generated on Folding@home. Conformational snapshots (frames) were stored at an interval of 1 ns/frame for subsequent analysis. The resulting final dataset contained 5000 trajectories, 1.1 ms of aggregate simulation time. Solute-only trajectories: The solute-only trajectories (with counterions) are available as MDTraj HDF5 files that contain both topology and trajectory information. A single trajectory (RUN0 CLONE0) (~42 MB) can be downloaded using the AWS CLI:

aws s3 --no-sign-request cp s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/munged/solute/17341/run0-clone0.h5 .

All HDF5 trajectories can be retrieved with

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/munged/solute/17341 .

Entire dataset: The raw Folding@home dataset is made available through the AWS Open Data Registry and can be retrieved through the AWS CLI. The dataset consists of a single project (PROJ17341) and has a RUN*/CLONE*/result* directory structure. RUNs denote different equilibrated starting structures. CLONEs denote different independent replica trajectories. To retrieve raw trajectory files in gromacs XTC format for the whole dataset, you can use the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/raw-data/PROJ17341 .

Folding@home initial files: System setup and input files can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/setup-files/17341 .

Contributors: Ivy Zhang, William G. Glass, Tristan I. Croll, Aoife M. Harbison, Elisa Fadda, John D. Chodera. License: All data is freely available under the Creative Commons CC0 (“No Rights Reserved”) license.

Input and Supporting Files: ---

Trajectory: Get Trajectory (102 GB)

Represented Proteins: spike RBD

Represented Structures: 7jx3

Models: SARS-CoV-2 spike receptor-binding domain bound with S309: ISOLDE refined model with N343 glycan

DESRES-ANTON-10897850 10 µs simulation of of the trimeric SARS-CoV-2 spike glycoprotein, no water or ions (10 µs )

D. E. Shaw Research

DESRES

A 10 µs simulation of the trimeric SARS-CoV-2 spike glycoprotein. System was initiated in a partially opened state (PDB entry 6VYB) which exhibited a high degree of conformational heterogeneity. In particular, the partially detached receptor binding domain sampled a variety of orientations, and further detached from the S2 fusion machinery. The simulation used the Amber ff99SB-ILDN force field for proteins, the CHARMM TIP3P model for water, and the generalized Amber force field for glycosylated asparagine. The C- and N-peptide termini, including those exposed due to missing loops in the published structural models, are capped with amide and acetyl groups respectively. The system was neutralized and salted with NaCl, with a final concentration of 0.15 M. The total number of atoms in the system was 715439 for the closed state. The interval between frames is 1.2 ns. The simulations were conducted at 310 K in the NPT ensemble. We have released new versions of these simulations with enhancements to the spike protein model in [DESRES-ANTON-11021566,11021571] (https://www.deshawresearch.com/downloads/download_trajectory_sarscov2.cgi/#DESRES-ANTON-11021566), since the one used in this simulation is incomplete in some of the disordered loop regions (i.e., resid 455 to 461, resid 469 to 488) and in glycan chains.

Input and Supporting Files:

DESRES-Trajectory_sarscov2-10897850-structure.tar.gz

DESRES-Trajectory_sarscov2-10897850.mp4

Trajectory: Get Trajectory (4.1 GB)

Represented Proteins: spike

Represented Structures: 6vyb

Models: Trimeric SARS-CoV-2 spike glycoprotein (open state) in aqueous solution

Riken CPR TMS, MD1_Down trajectory (1 microseconds )

Takaharu Mori, Jaewoon Jung, Chigusa Kobayashi, Hisham M. Dokainish, Suyong Re, Yuji Sugita

RIKEN CPR (Cluster for Pioneering Research), TMS (Theoretical molecular science) laboratory -- TMS (Theoretical molecular science) laboratory

The data set includes a trajectory file from Molecular Dynamics (MD) of a fully glycosylated SARS-CoV-2 S-protein in solution. Water molecules and counter ions were excluded. The starting structure is an inactive Down taken from CHARMM-GUI COVID-19 Archive (http://www.charmm-gui.org/docs/archive/covid19). We replaced counter ions K+ in the original model with Na+. The simulation used CHARMM36m force field for protein, and TIP3P water model. The simulation was performed using GENESIS. The coordinates were saved every 1 nanoseconds and aligned to S2 domain (Calpha atoms of residues 689-727, 854-1147).

Input and Supporting Files:

Down_pro-gly.psf

Trajectory: Get Trajectory (742MB)

Represented Proteins: spike

Represented Structures: 6vxx

Models:

Continuous trajectories of glycosylated SPIKE opening. (175 ns )

Amaro Lab and Chong Lab

All-atom MD trajectories from weighted ensemble simulations of glycosylated SPIKE protein, protein + glycans only. PSF, prmtop, DCDs, and WESTPA input files are provided. Starting structure based on model of the full-length spike in the closed state developed by the Amaro lab, which is modeled from 6VXX. Only the head region of the Spike was included in simulations from residues 16-1140.

Input and Supporting Files: ---

Trajectory: Get Trajectory (1.35 GB)

Represented Proteins: spike

Represented Structures: 6VXX

Models:

Simulations of SARS-CoV and SARS-CoV-2 RBD with ACE2 (2 µs )

Gumbart lab

Two-microsecond trajectories of the receptor-binding domains from SARS-CoV and SARS-CoV-2 spike protein bound to the human receptor, ACE2 (two replicas each). Simulation systems were constructed with VMD, equilibrated initially with NAMD, and then run for 2 µs each with Amber16. Simulations used a 4-fs timestep enabled by hydrogen-mass repartitioning (HMR).

Input and Supporting Files: ---

Trajectory: Get Trajectory (5.3 GB)

Represented Proteins: spike RBD ACE2

Represented Structures: 2AJF 6M17

Models: ---

Trajectories of full-length SPIKE protein in the Closed state. (1.7 µs )

Amaro Lab

All-atom MD simulations of full-length SPIKE protein in the Closed state, protein + glycans only (not aligned). PSF and DCDs files are provided.

Input and Supporting Files: ---

Trajectory: Get Trajectory (13 GB)

Represented Proteins: spike

Represented Structures: 6VXX

Models:

Cluster ensemble of Intermediate 3a (500 ns )

sugita lab

CPR

30 PDB structures of the intermediate (I3a) cluster obtained from gREST_SSCR simulations starting from Down conformation. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (30.8 MB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Trimeric SARS-CoV-2 spike glycoprotein (Down state) with and without simulation box

MD simulations of trimeric SARS-Cov2 spike protein ectodomain in explicit solvent. Data were collected for apo, linoleic acid bound and other putative ligands (3x200 ns in each case) (24 x 200 ns trajectories (solvent removed) )

Deborah K Shoemark

University of Bristol, UK -- BrisSynBio and Mulholland

The CryoEM stuctures of the apo and linoleic acid bound SARS-Cov2 spike protein trimer (residues 15/25 to 1139) were used to build complete atomistic models. Other putative ligands, including cholesterol and vitamins, retinoids and steroids identified by docking with BUDE, were simulated in both open and closed states. The closed and open structures have 42 and 43 disulfide bonds respectively. Simulations were performed with GROMACS 2019.x. the file Spike_MD_simulations.tgz contains:

• Spike_MD_simulations/
• Spike_MD_simulations/WT_closed-SARS2-spike_apo/
• Spike_MD_simulations/WT_closed-SARS2-spike_apo/01_WT_closed_apo_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_apo/02_WT_closed_apo_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_apo/03_WT_closed_apo_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_apo/01_WT_closed_apo_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_apo/02_WT_closed_apo_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_apo/03_WT_closed_apo_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_apo/README
• Spike_MD_simulations/WT_closed-SARS2-spike_cholesterol/
• Spike_MD_simulations/WT_closed-SARS2-spike_cholesterol/01_WT_closed-OK_CLR_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_cholesterol/02_WT_closed-OK_CLR_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_cholesterol/03_WT_closed-OK_CLR_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_cholesterol/01_WT_closed-OK_CLR_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_cholesterol/02_WT_closed-OK_CLR_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_cholesterol/03_WT_closed-OK_CLR_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_cholesterol/README
• Spike_MD_simulations/WT_closed-SARS2-spike_dexamethasone/
• Spike_MD_simulations/WT_closed-SARS2-spike_dexamethasone/01_clean_WT_closed_dexys_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_dexamethasone/02_clean_WT_closed_dexys_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_dexamethasone/03_clean_WT_closed_dexys_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_dexamethasone/README
• Spike_MD_simulations/WT_closed-SARS2-spike_dexamethasone/01_clean_WT_closed_dexys_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_dexamethasone/02_clean_WT_closed_dexys_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_dexamethasone/03_clean_WT_closed_dexys_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_LA/
• Spike_MD_simulations/WT_closed-SARS2-spike_LA/01_WT_closed_LA_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_LA/02_WT_closed_LA_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_LA/03_WT_closed_LA_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_LA/01_WT_closed_LA_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_LA/02_WT_closed_LA_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_LA/03_WT_closed_LA_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_LA/README
• Spike_MD_simulations/WT_open-SARS2-spike_apo/
• Spike_MD_simulations/WT_open-SARS2-spike_apo/01_WT-OK_open_apo_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_apo/02_WT-OK_open_apo_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_apo/03_WT-OK_open_apo_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_apo/README
• Spike_MD_simulations/WT_open-SARS2-spike_apo/01_WT-OK_open_apo_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_apo/02_WT-OK_open_apo_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_apo/03_WT-OK_open_apo_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_cholesterol/
• Spike_MD_simulations/WT_open-SARS2-spike_cholesterol/01_WT-open-OK_CLR_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_cholesterol/01_WT-open-OK_CLR_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_cholesterol/02_WT-open-OK_CLR_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_cholesterol/02_WT-open-OK_CLR_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_cholesterol/03_WT-open-OK_CLR_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_cholesterol/03_WT-open-OK_CLR_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_cholesterol/README
• Spike_MD_simulations/WT_open-SARS2-spike_dexamethasone/
• Spike_MD_simulations/WT_open-SARS2-spike_dexamethasone/01_clean_WT_open_dexys_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_dexamethasone/03_clean_WT_open_dexys_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_dexamethasone/01_clean_WT_open_dexys_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_dexamethasone/03_clean_WT_open_dexys_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_dexamethasone/02_clean_WT_open_dexys_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_dexamethasone/02_clean_WT_open_dexys_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_dexamethasone/README
• Spike_MD_simulations/WT_open-SARS2-spike_LA/
• Spike_MD_simulations/WT_open-SARS2-spike_LA/01_WT-OK_open_LAs_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_LA/02_WT-OK_open_LAs_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_LA/03_WT-OK_open_LAs_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_LA/01_WT-OK_open_LAs_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_LA/02_WT-OK_open_LAs_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_LA/03_WT-OK_open_LAs_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_LA/README
• Spike_MD_simulations/README

Input and Supporting Files:

Spike_MD_simulations.tgz

Trajectory: Get Trajectory (9 GB)

Represented Proteins: spike

Represented Structures: 6ZB5

Models: ---

Folding@home simulations of the SARS-CoV-2 spike RBD with N501Y mutation bound to human ACE2 (953.7 µs )

Ivy Zhang

Folding@home -- Chodera lab

All-atom MD simulations of the SARS-CoV-2 spike protein receptor binding domain (RBD) with N501Y mutation bound to human angiotensin converting enzyme-related carboypeptidase (ACE2), simulated using Folding@Home. Complete details of this simulation are available here. Brief details appear below. Publication: https://doi.org/10.1016/j.cell.2021.01.037 System preparation: The RBD:ACE2 complex was constructed from individual RBD (PDB: 6m0j, Chain E) and ACE2 (PDB: 1r42, Chain A) monomers aligned to the full RBD:ACE2 structure (PDB: 6m0j. These structural models were further refined by Tristan Croll using ISOLDE (Croll, 2018) and deposited in the Coronavirus Structural Taskforce (CST) database (Croll et al., 2020) to produce refined 6m0j and refined 1r42 models. The RBD N501 was mutated to TYR using PyMOL 2.3.2. The resulting RBD and ACE2 monomers were then aligned in PyMOL 2.3.2 to the CST 6m0j structure to create an initial RBD:ACE2 complex. Full glycosylation patterns for ACE2 and RBD glycans were determined from Shajahan et al. For the constructed RBD:ACE2 complex, these included sites: N53, N90, N103, N322, N432, N546, and N690 on ACE2 and N343 on the RBD. Base NAG residues of each glycan structure (FA2, FA26G1, FA2, FA2, FA2G2, A2, FA2, FA2G2, respectively) were acquired from Elisa Fadda. Each glycan was then aligned to the corresponding NAG stub in the RBD:ACE2 model in and any resulting clashes were refined in ISOLDE. Full details of the glycosylation patterns / structures used and full workflow are available here. Folding@home simulation: The equilibrated structure was then used to initiate parallel distributed MD simulations on Folding@home (Shirts and Pande, 2000, Zimmerman et al., 2020). Simulations were run with OpenMM 7.4.2 (Folding@home core22 0.0.13). Production simulations used the same Langevin integrator as the NpT equilibration described above. In total, 5000 independent MD simulations were generated on Folding@home. Conformational snapshots (frames) were stored at an interval of 1 ns/frame for subsequent analysis. The resulting final dataset contained 5000 trajectories and 953.7 µs of aggregate simulation time. Solute-only trajectories: The solute-only trajectories (with counterions) are available as MDTraj HDF5 files that contain both topology and trajectory information. A single trajectory of the WT RBD (RUN3) can be downloaded using the AWS CLI:

aws s3 --no-sign-request cp s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-ACE2-RBD/munged/solute/17344/run0-clone0.h5 .

All HDF5 trajectories can be retrieved with

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-ACE2-RBD/munged/solute/17344 .

Entire dataset: The raw Folding@home dataset is made available through the AWS Open Data Registry and can be retrieved through the AWS CLI. The dataset consists of a single project (PROJ17344) and has a RUN*/CLONE*/result* directory structure. RUNs denote different equilibrated starting structures. CLONEs denote different independent replica trajectories. To retrieve raw trajectory files in gromacs XTC format for the whole dataset, you can use the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-ACE2-RBD/raw-data/PROJ17344 .

Folding@home initial files: System setup and input files can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-ACE2-RBD/setup/17344 .

Contributors: Ivy Zhang, William G. Glass, Tristan I. Croll, Aoife M. Harbison, Elisa Fadda, John D. Chodera. License: All data is freely available under the Creative Commons CC0 (“No Rights Reserved”) license.

Input and Supporting Files: ---

Trajectory: Get Trajectory (132 GB)

Represented Proteins: spike RBD ACE2

Represented Structures: 6m0j 1R42

Models: SARS-CoV-2 spike receptor-binding domain: ISOLDE refined model with N343 glycan and N501Y mutation Native Human Angiotensin Converting Enzyme-Related Carboxypeptidase (ACE2): ISOLDE refined model with glycans

Riken CPR TMS, TMD2_toDown trajectory (20 nanoseconds )

Takaharu Mori, Jaewoon Jung, Chigusa Kobayashi, Hisham M. Dokainish, Suyong Re, Yuji Sugita

RIKEN CPR (Cluster for Pioneering Research), TMS (Theoretical molecular science) laboratory -- TMS (Theoretical molecular science) laboratory

The data set includes a trajectory file from Targeted Molecular Dynamics (TMD) simulations of a fully glycosylated SARS-CoV-2 S-protein in solution. Water molecules and counter ions were excluded. The data includes trajectory from TMD simulation of Up to Down forms. The simulations used CHARMM36m force field for protein, and TIP3P water model. The simulations were performed using GENESIS. The coordinates were saved every 1 nanoseconds and aligned to S2 domain (Calpha atoms of residues 689-727, 854-1147).

Input and Supporting Files:

Up_pro-gly.psf

Trajectory: Get Trajectory (15MB)

Represented Proteins: spike

Represented Structures: 6vsb 6vxx

Models:

Riken CPR TMS, TMD3_toUp trajectory (50 nanoseconds )

Takaharu Mori, Jaewoon Jung, Chigusa Kobayashi, Hisham M. Dokainish, Suyong Re, Yuji Sugita

RIKEN CPR (Cluster for Pioneering Research), TMS (Theoretical molecular science) laboratory -- TMS (Theoretical molecular science) laboratory

The data set includes a trajectory file from Targeted Molecular Dynamics (TMD) simulations of a fully glycosylated SARS-CoV-2 S-protein in solution. Water molecules and counter ions were excluded. The data includes trajectory from TMD simulation of Down to Up forms. The simulations used CHARMM36m force field for protein, and TIP3P water model. The simulations were performed using GENESIS. The coordinates were saved every 1 nanoseconds and aligned to S2 domain (Calpha atoms of residues 689-727, 854-1147).

Input and Supporting Files:

Down_pro-gly.psf

Trajectory: Get Trajectory (38MB)

Represented Proteins: spike

Represented Structures: 6vxx 6vsb

Models:

Folding@home simulations of the SARS-CoV-2 spike RBD bound to human ACE2 (725.3 µs )

Ivy Zhang

Folding@home -- Chodera lab

All-atom MD simulations of the SARS-CoV-2 spike protein receptor binding domain (RBD) bound to human angiotensin converting enzyme-related carboypeptidase (ACE2), simulated using Folding@Home. The “wild-type” RBD and three mutants (N439K, K417V, and the double mutant N439K/K417V) were simulated.

Complete details of this simulation are available here. Brief details appear below.

Publication: https://doi.org/10.1016/j.cell.2021.01.037

System preparation: The RBD:ACE2 complex was constructed from individual RBD (PDB: 6m0j, Chain E) and ACE2 (PDB: 1r42, Chain A) monomers aligned to the full RBD:ACE2 structure (PDB: 6m0j. These structural models were further refined by Tristan Croll using ISOLDE (Croll, 2018) and deposited in the Coronavirus Structural Taskforce (CST) database (Croll et al., 2020) to produce refined 6m0j and refined 1r42 models. The resulting RBD and ACE2 monomers were then aligned in PyMOL 2.3.2 to the CST 6m0j structure to create an initial RBD:ACE2 complex.

Full glycosylation patterns for ACE2 and RBD glycans were determined from Shajahan et al. For the constructed RBD:ACE2 complex, these included sites: N53, N90, N103, N322, N432, N546, and N690 on ACE2 and N343 on the RBD. Base NAG residues of each glycan structure (FA2, FA26G1, FA2, FA2, FA2G2, A2, FA2, FA2G2, respectively) were acquired from Elisa Fadda. Each glycan was then aligned to the corresponding NAG stub in the RBD:ACE2 model in and any resulting clashes were refined in ISOLDE. Full details of the glycosylation patterns / structures used and full workflow are available here.

Folding@home simulation: The equilibrated structure was then used to initiate parallel distributed MD simulations on Folding@home (Shirts and Pande, 2000, Zimmerman et al., 2020). Simulations were run with OpenMM 7.4.2 (Folding@home core22 0.0.13). Production simulations used the same Langevin integrator as the NpT equilibration described above. In total, 8000 independent MD simulations were generated on Folding@home. Conformational snapshots (frames) were stored at an interval of 0.5 ns/frame for subsequent analysis. The resulting final dataset contained 8000 trajectories, 725.3 us of aggregate simulation time, and 1450520 frames. Solute-only trajectories: The solute-only trajectories (with counterions) are available as MDTraj HDF5 files that contain both topology and trajectory information. A single trajectory of the WT RBD (RUN3) (~30 MB) can be downloaded using the AWS CLI:

aws s3 --no-sign-request cp s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-ACE2-RBD/munged/solute/PROJ17311/run3-clone0.h5 .

All HDF5 trajectories (~300 GB) can be retrieved with

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-ACE2-RBD/munged/solute/PROJ17311 .

Entire dataset: The raw Folding@home dataset is made available through the AWS Open Data Registry and can be retrieved through the AWS CLI. The dataset consists of a single project (PROJ17311) and has a RUN*/CLONE*/result* directory structure. RUNs denote different RBD mutants: N439K (RUN0), K417V (RUN1), N439K/K417V (RUN2), and WT (RUN3). CLONEs denote different independent replica trajectories.

To retrieve raw trajectory files in gromacs XTC format for the whole dataset (7 TB), you can use the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-ACE2-RBD/raw-data/PROJ17311 .

Folding@home initial files: System setup and input files can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-ACE2-RBD/setup/PROJ17311 .

Contributors: Ivy Zhang, William G. Glass, Tristan I. Croll, Aoife M. Harbison, Elisa Fadda, John D. Chodera.

License: All data is freely available under the Creative Commons CC0 (“No Rights Reserved”) license.

Input and Supporting Files: ---

Trajectory: Get Trajectory (341 GB)

Represented Proteins: spike RBD ACE2

Represented Structures: 6m0j 1R42

Models: SARS-CoV-2 spike receptor-binding domain: ISOLDE refined model with N343 glycan Native Human Angiotensin Converting Enzyme-Related Carboxypeptidase (ACE2): ISOLDE refined model with glycans

Riken CPR TMS, MD2_Down trajectory (200 nanoseconds )

Takaharu Mori, Jaewoon Jung, Chigusa Kobayashi, Hisham M. Dokainish, Suyong Re, Yuji Sugita

RIKEN CPR (Cluster for Pioneering Research), TMS (Theoretical molecular science) laboratory -- TMS (Theoretical molecular science) laboratory

The data set includes a trajectory file from Molecular Dynamics (MD) of a fully glycosylated SARS-CoV-2 S-protein in solution. Water molecules and counter ions were excluded. The starting structure is an inactive Down taken from CHARMM-GUI COVID-19 Archive (http://www.charmm-gui.org/docs/archive/covid19). We replaced counter ions K+ in the original model with Na+. The simulation used CHARMM36m force field for protein, and TIP3P water model. The simulation was performed using GENESIS. The coordinates were saved every 1 nanoseconds and aligned to S2 domain (Calpha atoms of residues 689-727, 854-1147).

Input and Supporting Files:

Down_pro-gly.psf

Trajectory: Get Trajectory (149MB)

Represented Proteins: spike

Represented Structures: 6vxx

Models:

Gromacs 60 ns MD of SARS-CoV-2 spike trimer, All Atom model (60 ns )

Dmitry Morozov

University of Jyvaskyla

This trajectory is from a 60 ns MD simulation of the SARS-CoV-2 spike protein. The protein was solvated in a 20 x 20 x 20 nm water box containing 0.1 M NaCl. The simulation was performed with Gromacs 2018.8 on the Puhti cluster located at the CSC-IT using the Charmm27 force field. The interval between frames is 80 ps. The simulation was conducted in the NPT ensemble (1 bar). This trajectory is all atom.

Input and Supporting Files:

trimer

Trajectory: Get Trajectory (2.0 GB)

Represented Proteins: spike

Represented Structures: 6VXX

Models: SARS-CoV-2 spike protein trimer (closed state) model for MD simulations

DESRES-ANTON-11021571 10 µs simulation of of the trimeric SARS-CoV-2 spike glycoprotein in aqueous solution (10 µs )

D. E. Shaw Research

DESRES

10 µs simulation trajectory of the trimeric SARS-CoV-2 spike glycoprotein with additional loop structures and glycan chains to improve the spike protein model originally released in DESRES-ANTON-[10897136,10897850]. Trajectory was initiated in a partially opened state (PDB entry 6VYB). The simulation used the Amber ff99SB-ILDN force field for proteins, the CHARMM TIP3P model for water, and the generalized Amber force field for glycosylated asparagine. The C- and N-peptide termini are capped with amide and acetyl groups respectively. The system was neutralized and salted with NaCl, with a final concentration of 0.15 M. The interval between frames is 1.2 ns. The simulations were conducted at 310 K in the NPT ensemble.

Input and Supporting Files:

DESRES-Trajectory_sarscov2-11021571-structure.tar.gz

DESRES-Trajectory_sarscov2-11021571.mp4

Trajectory: Get Trajectory (67 GB)

Represented Proteins: spike

Represented Structures: 6vyb

Models: Trimeric SARS-CoV-2 spike glycoprotein (open state) in aqueous solution

Cluster ensemble of 1UP/open conformations (300 ns )

sugita lab

CPR

30 PDB structures of the 1Up/open conformations obtained from gREST_SSCR simulations starting from Up conformation. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (30.8 MB)

Represented Proteins: spike

Represented Structures: 6vyb

Models: Trimeric SARS-CoV-2 spike glycoprotein (1Up state) with and without simulation box

SIRAH-CoV2 initiative - S2 Spike core fragment in postfusion state (10 µs )

Florencia Klein

Institut Pasteur de Montevideo -- Biomolecular Simulations Laboratory

This dataset contains the trajectory of a 10 microseconds-long coarse-grained molecular dynamics simulation of SARS-CoV2 Spike S2 fragment in its postfusion form (PDB id: 6M1V). Simulations have been performed using the SIRAH force field running with the Amber18 package at the Uruguayan National Center for Supercomputing (ClusterUY) under the conditions reported in Machado et al. JCTC 2019, adding 150 mM NaCl according to Machado & Pantano JCTC 2020.

The files 6M1V_SIRAHcg_rawdata_0-5us.tar, and 6M1V_SIRAHcg_rawdata_5-10us.tar contain all the raw information required to visualize (on VMD 1.9.3), analyze, backmap, and eventually continue the simulations using Amber18 or higher. Step-By-Step tutorials for running, visualizing, and analyzing CG trajectories using SirahTools can be found at SIRAH website.

Additionally, the file 6M1V_SIRAHcg_10us_prot.tar contains only the protein coordinates, while 6M1V_SIRAHcg_10us_prot_skip10ns.tar contains one frame every 10ns.

To take a quick look at the trajectory:

1- Untar the file 6M1V_SIRAHcg_10us_prot_skip10ns.tar

2- Open the trajectory on VMD using the command line: vmd 6M1V_SIRAHcg_prot.prmtop 6M1V_SIRAHcg_prot_10us_skip10ns.ncrst 6M1V_SIRAHcg_prot_10us_skip10ns.nc -e sirah_vmdtk.tcl

Note that you can use normal VMD drawing methods as vdw, licorice, etc., and coloring by restype, element, name, etc.

Input and Supporting Files: ---

Trajectory: Get Trajectory (17.8 GB)

Represented Proteins: spike S2

Represented Structures: 6M1V

Models: ---

SIRAH-CoV2 initiative - Spike´s RBD/ACE2-B0AT1 complex (4 µs )

Florencia Klein

Institut Pasteur de Montevideo -- Biomolecular Simulations Laboratory

This dataset contains an trajectory of four microseconds-long coarse-grained molecular dynamics simulation of the hexameric complex between SARS-CoV2 Spike´s RBD, ACE2, and B0AT1 (PDB id: 6M17). Simulations have been performed using the SIRAH force field running with the Amber18 package at the Uruguayan National Center for Supercomputing (ClusterUY) under the conditions reported in Machado et al. JCTC 2019, adding 150 mM NaCl according to Machado & Pantano JCTC 2020. Zinc ions were parameterized as reported in Klein et al. 2020.

The files 6M17_SIRAHcg_rawdata_0-1.tar, 6M17_SIRAHcg_rawdata_1-2.tar, 6M17_SIRAHcg_rawdata_2-3.tar, and 6M17_SIRAHcg_rawdata_3-4.tar, contain all the raw information required to visualize (on VMD), analyze, backmap, and eventually continue the simulations using Amber18 or higher. Step-By-Step tutorials for running, visualizing, and analyzing CG trajectories using SirahTools can be found at SIRAH website.

Additionally, the file 6M17_SIRAHcg_4us_prot.tar contains only the protein coordinates, while 6M17_SIRAHcg_4us_prot_skip10ns.tar contains one frame every 10ns.

To take a quick look at the trajectory:

1- Untar the file 6M17_SIRAHcg_4us_prot_skip10ns.tar

2- Open the trajectory on VMD using the command line: vmd 6M17_SIRAHcg_prot.prmtop 6M17_SIRAHcg_prot_4us_skip10ns.ncrst 6M17_SIRAHcg_prot_4us_skip10ns.nc -e sirah_vmdtk.tcl

Note that you can use normal VMD drawing methods as vdw, licorice, etc., and coloring by restype, element, name, etc.

Input and Supporting Files: ---

Trajectory: Get Trajectory (20.1 GB)

Represented Proteins: spike RBD ACE2 BoAT1

Represented Structures: 6M17

Models: ---

DESRES-ANTON-10906555 2 µs simulations of 50 FDA approved or investigational drug molecules binding to a construct of the SARS-CoV-2 trimeric spike protein, no water or ions (2 µs )

D. E. Shaw Research

DESRES

50 2 µs trajectories of FDA approved or investigational drug molecules that in simulation remained bound to a construct of the SARS-CoV-2 trimeric spike protein at positions that might conceivably allosterically disrupt the interaction between these proteins. The small molecule drugs and their initial binding poses were chosen from a combination of molecular dynamics simulation and docking performed using an FDA-investigational drug library. The 50 putative spike protein binding small molecules located at three regions on the spike trimer, a pocket in the RBD whose formation may possibly enhance RBD-RBD interactions in the closed conformation (8 molecules), a pocket between the two RBDs in the closed conformation (29 molecules), and a pocket that involves three RBDs in the closed conformation (13 molecules). The simulations used the Amber ff99SB-ILDN force field for proteins, the CHARMM TIP3P model for water, and the generalized Amber force field for small molecules. The C- and N-peptide termini were capped with amide and acetyl groups respectively. The spike trimer construct was modeled from PDB entries 6VXX and 6VW1, only retaining the RBD and a short region from S1 fusion protein as a minimal system for maintaining a trimer assembly. The system was neutralized and salted with NaCl, with a final concentration of 0.15 M. The interval between frames is 1.2 ns. The simulations were conducted at 310 K in the NPT ensemble.

Input and Supporting Files:

DESRES-Trajectory_sarscov2-10906555-set_spike-structure.tar.gz

DESRES-Trajectory_sarscov2-10906555-set_spike-table.csv

DESRES-Trajectory_sarscov2-10906555.mp4

Trajectory: Get Trajectory (14 GB)

Represented Proteins: spike RBD

Represented Structures: 6vw1 6vxx

Models: SARS-CoV-2 trimeric spike protein binding to FDA approved or investigational drug molecules

Folding@home simulations of the SARS-CoV-2 spike RBD with P337L mutation bound to monoclonal antibody S309 (923.2 µs )

Ivy Zhang

Folding@home -- Chodera lab

All-atom MD simulations of the SARS-CoV-2 spike protein receptor binding domain (RBD) with P337L mutation bound to monoclonal antibody S309, simulated using Folding@Home. Complete details of this simulation are available here. Brief details appear below. Publication: https://doi.org/10.1038/s41586-021-03807-6 System preparation: The RBD:S309 complex was constructed from PDB ID 7JX3 (Chains A, B, and R). 7JX3 was first refined using ISOLDE to better fit the experimental electron density using detailed manual inspection. Refinement included adjusting several rotamers, flipping several peptide bonds, fixing several weakly resolved waters, and building in a missing four-residue-long loop. Though the N343 glycan N-Acetylglucosamine (NAG) was present in 7JX3, ISOLDE was used to construct a complex glycan at N343. The full glycosylation pattern was determined from Shajahan et al. and Watanabe et al. The glycan structure used for N343 (FA2G2) corresponds to the most stable conformer obtained from multi microsecond molecular dynamics (MD) simulations of cumulative sampling. The base NAG residue in FA2G2 was aligned to the corresponding NAG stub in the RBD:S309 model and any resulting clashes were refined in ISOLDE. PyMOL was used to mutate RBD’s P337 to LEU. The equilibrated structure was then used to initiate parallel distributed MD simulations on Folding@home (Shirts and Pande, 2000, Zimmerman et al., 2020). Simulations were run with OpenMM 7.4.2 (Folding@home core22 0.0.13). Production simulations used the same Langevin integrator as the NPT equilibration described above. In total, 5985 independent MD simulations were generated on Folding@home. Conformational snapshots (frames) were stored at an interval of 1 ns/frame for subsequent analysis. The resulting final dataset contained 5985 trajectories, 923.2 µs of aggregate simulation time. Solute-only trajectories: The solute-only trajectories (with counterions) are available as MDTraj HDF5 files that contain both topology and trajectory information. A single trajectory (RUN0 CLONE0) can be downloaded using the AWS CLI:

aws s3 --no-sign-request cp s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/munged/solute/17343/run0-clone0.h5 .

All HDF5 trajectories can be retrieved with

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/munged/solute/17343 .

Entire dataset: The raw Folding@home dataset is made available through the AWS Open Data Registry and can be retrieved through the AWS CLI. The dataset consists of a single project (PROJ17343) and has a RUN*/CLONE*/result* directory structure. RUNs denote different equilibrated starting structures. CLONEs denote different independent replica trajectories. To retrieve raw trajectory files in gromacs XTC format for the whole dataset, you can use the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/raw-data/PROJ17343 .

Folding@home initial files: System setup and input files can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/setup-files/17343 .

Contributors: Ivy Zhang, William G. Glass, Tristan I. Croll, Aoife M. Harbison, Elisa Fadda, John D. Chodera. License: All data is freely available under the Creative Commons CC0 (“No Rights Reserved”) license.

Input and Supporting Files: ---

Trajectory: Get Trajectory (91 GB)

Represented Proteins: spike RBD

Represented Structures: 7jx3

Models: SARS-CoV-2 spike receptor-binding domain bound with S309: ISOLDE refined model with N343 glycan and P337L mutation

Cluster ensemble of 1UP top populated cluster (300 ns )

sugita lab

CPR

30 PDB structures of the top populated cluster obtained from gREST_SSCR simulations starting from 1Up conformation. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (30.8 MB)

Represented Proteins: spike

Represented Structures: 6vyb

Models: Trimeric SARS-CoV-2 spike glycoprotein (1Up state) with and without simulation box

Clusters center of gREST from Down State simulations (500 ns )

sugita lab

CPR

PDB of cluster centers representing 13 clusters obtained from gREST_SSCR simulations starting from Down conformation. This includes Down symmetric (D1_Sym.pdb and D2_Sym.pdb), Down asymmetric (D1_asym.pdb and D2_asym.pdb), Intermediate 1 (I1a.pdb, I1b.pdb and I1c.pdb), Intermediate 2 (I2a.pdb, I2b.pdb and I2c.pdb), Intermediate 3 (I3a.pdb and I3b.pdb) and 1Up like (1U_L.pdb) conformations. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (13.3 MB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Trimeric SARS-CoV-2 spike glycoprotein (Down state) with and without simulation box

Cluster ensemble of Intermediate 2a (500 ns )

sugita lab

CPR

30 PDB structures of the intermediate (I2a) cluster obtained from gREST_SSCR simulations starting from Down conformation. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (30.8 MB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Trimeric SARS-CoV-2 spike glycoprotein (Down state) with and without simulation box

Riken CPR TMS, TMD1_toUp trajectory (20 nanoseconds )

Takaharu Mori, Jaewoon Jung, Chigusa Kobayashi, Hisham M. Dokainish, Suyong Re, Yuji Sugita

RIKEN CPR (Cluster for Pioneering Research), TMS (Theoretical molecular science) laboratory -- TMS (Theoretical molecular science) laboratory

The data set includes a trajectory file from Targeted Molecular Dynamics (TMD) simulations of a fully glycosylated SARS-CoV-2 S-protein in solution. Water molecules and counter ions were excluded. The data includes trajectory from TMD simulation of Down to Up forms. The simulations used CHARMM36m force field for protein, and TIP3P water model. The simulations were performed using GENESIS. The coordinates were saved every 1 nanoseconds and aligned to S2 domain (Calpha atoms of residues 689-727, 854-1147).

Input and Supporting Files:

Down_pro-gly.psf

Trajectory: Get Trajectory (15MB)

Represented Proteins: spike

Represented Structures: 6vxx 6vsb

Models:

Cluster ensemble of Down asymmetric (500 ns )

sugita lab

CPR

30 PDB structures of the Down asymmetric (D1_asym) cluster obtained from gREST_SSCR simulations starting from Down conformation. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (30.8 MB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Trimeric SARS-CoV-2 spike glycoprotein (Down state) with and without simulation box

Trajectory of the Spike protein in complex with human ACE2 (50 ns )

Oostenbrink Lab

University of Natural Resources and Life Sciences, Vienna

Atomistic MD simulations of the Spike protein in complex with the human ACE2 receptor, most probale glycosylations are added.

Input and Supporting Files:

inputdata.tar.gz

Trajectory: Get Trajectory (43 GB)

Represented Proteins: spike ACE2

Represented Structures: 6vyb 6m17

Models: Spike protein in complex with human ACE2

1 microsecond trajecotry of glycosylated spike protein in open state for pdb:6VSB embedded in viral membrane (1 µs )

Klauda lab

All atom simulation of full-glycosylated spike protein in open state (pdb:6VSB) embedded in viral membrane. The structure was taken from Charmm-Gui at http://www.charmm-gui.org/?doc=archive&lib=covid19 where 8 models were built for the open state. For MD simulations we used model 1-2-1 provided by Im et. al. The PSF, PDB and XTC files are uploaded

Input and Supporting Files: ---

Trajectory: Get Trajectory (12 GB)

Represented Proteins: spike

Represented Structures: 6VSB

Models:

Cluster ensemble of Down symmetric (500 ns )

sugita lab

CPR

30 PDB structures of the Down symmetric (D1_Sym) cluster obtained from gREST_SSCR simulations starting from Down conformation. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (30.8 MB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Trimeric SARS-CoV-2 spike glycoprotein (Down state) with and without simulation box

Folding@home simulations of the SARS-CoV-2 spike RBD bound to monoclonal antibody S2H97 (623.7 us )

Ivy Zhang

Folding@home -- Chodera lab

All-atom MD simulations of the SARS-CoV-2 spike protein receptor binding domain (RBD) bound to monoclonal antibody S2H97, simulated using Folding@Home. Complete details of this simulation are available here. Brief details appear below. Publication: https://doi.org/10.1038/s41586-021-03807-6 System preparation: The RBD:S2H97 complex was constructed from PDB ID 7M7W (Chains S, C, and D). 7M7W was first refined using ISOLDE to better fit the experimental electron density using detailed manual inspection. Refinement included building in a missing four-residue-long loop. ISOLDE was used to construct a complex glycan at N343. The full glycosylation pattern was determined from Shajahan et al. and Watanabe et al. The glycan structure used for N343 (FA2G2) corresponds to the most stable conformer obtained from multi microsecond molecular dynamics (MD) simulations of cumulative sampling. The equilibrated structure was then used to initiate parallel distributed MD simulations on Folding@home (Shirts and Pande, 2000, Zimmerman et al., 2020). Simulations were run with OpenMM 7.4.2 (Folding@home core22 0.0.13). Production simulations used the same Langevin integrator as the NpT equilibration described above. In total, 4985 independent MD simulations were generated on Folding@home. Conformational snapshots (frames) were stored at an interval of 1 ns/frame for subsequent analysis. The resulting final dataset contained 4985 trajectories, 623.7 us of aggregate simulation time. Solute-only trajectories: The solute-only trajectories (with counterions) are available as MDTraj HDF5 files that contain both topology and trajectory information. A single trajectory (RUN0 CLONE0) (~29 MB) can be downloaded using the AWS CLI:

aws s3 --no-sign-request cp s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/munged/solute/17347/run0-clone0.h5 .

All HDF5 trajectories can be retrieved with

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/munged/solute/17347 .

Entire dataset: The raw Folding@home dataset is made available through the AWS Open Data Registry and can be retrieved through the AWS CLI. The dataset consists of a single project (PROJ17347) and has a RUN*/CLONE*/result* directory structure. RUNs denote different equilibrated starting structures. CLONEs denote different independent replica trajectories. To retrieve raw trajectory files in gromacs XTC format for the whole dataset, you can use the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/raw-data/PROJ17347 .

Folding@home initial files: System setup and input files can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/setup-files/17347 .

Contributors: Ivy Zhang, William G. Glass, Tristan I. Croll, Aoife M. Harbison, Elisa Fadda, John D. Chodera. License: All data is freely available under the Creative Commons CC0 (“No Rights Reserved”) license.

Input and Supporting Files: ---

Trajectory: Get Trajectory (60 GB)

Represented Proteins: spike RBD

Represented Structures: 7m7w

Models: SARS-CoV-2 spike receptor-binding domain bound with S2H97: ISOLDE refined model with N343 glycan

Trajectories of full-length SPIKE protein in the Open state (N165A / N234A mutations). (4.2 µs )

Amaro Lab

All-atom MD simulations of full-length SPIKE protein in the Open state bearing N165A and N234A mutations, protein + glycans only (not aligned). PSF and DCDs files are provided.

Input and Supporting Files: ---

Trajectory: Get Trajectory (31 GB)

Represented Proteins: spike

Represented Structures: 6VSB

Models:

PMF calculations of SARS-CoV-2 spike opening

Gumbart lab

Conformations (~500) along the opening paths of the SARS-CoV-2 spike trimer with and without glycans as well as with the diproline mutation. Simulation systems were constructed with VMD, equilibrated initially with NAMD, and then used for two-dimensional replica-exchange umbrella sampling. Conformations provided here are taken from the minimum free-energy path between 1-RBD up and down states in each potential of mean force (PMF). Note that each DCD does not represent a continuous simulation trajecotry. Simulations used a 4-fs timestep enabled by hydrogen-mass repartitioning (HMR).

Input and Supporting Files: ---

Trajectory: Get Trajectory (962 MB)

Represented Proteins: spike RBD ACE2

Represented Structures: 6VYB 6XR8

Models: ---

Nonequilibrium simulations of the SARS-Cov-2 wild-type and D614G spike (180 replicates, 5 ns each )

A.S.F. Oliveira

University of Bristol -- Mulholland Lab

Nonequilibrium MD simulation of the unglycosylated and uncleaved ectodomain of the SARS-CoV-2 wild-type and D614G spike

Input and Supporting Files:

nonequilibrium_simulations.tar.gz

Trajectory: Get Trajectory (23 GB)

Represented Proteins: spike

Represented Structures: https://www.rcsb.org/structure/6ZB5

Models: ---

DESRES-ANTON-11021566 10 µs simulation of of the trimeric SARS-CoV-2 spike glycoprotein in aqueous solution (10 µs )

D. E. Shaw Research

DESRES

10 µs simulation trajectory of the trimeric SARS-CoV-2 spike glycoprotein with additional loop structures and glycan chains to improve the spike protein model originally released in DESRES-ANTON-[10897136,10897850]. Trajectory was initiated in the closed state (PDB entry 6VXX). The simulation used the Amber ff99SB-ILDN force field for proteins, the CHARMM TIP3P model for water, and the generalized Amber force field for glycosylated asparagine. The C- and N-peptide termini are capped with amide and acetyl groups respectively. The system was neutralized and salted with NaCl, with a final concentration of 0.15 M. The interval between frames is 1.2 ns. The simulations were conducted at 310 K in the NPT ensemble.

Input and Supporting Files:

DESRES-Trajectory_sarscov2-11021566-structure.tar.gz

DESRES-Trajectory_sarscov2-11021566.mp4

Trajectory: Get Trajectory (51 GB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Improved trimeric SARS-CoV-2 spike glycoprotein (closed state) in aqueous solution

Folding@home simulations of the apo SARS-CoV-2 spike RBD (without glycosylation) (1.9 ms )

Ivy Zhang

Folding@home -- Chodera lab

All-atom MD simulations of the SARS-CoV-2 spike protein receptor binding domain (RBD) (without glycosylation), simulated using Folding@Home. Complete details of this simulation are available here. Brief details appear below. Publication: https://doi.org/10.1016/j.cell.2021.01.037 System preparation: The RBD complex was constructed from PDB ID 6M0J (Chain B). 6M0J was refined using ISOLDE to better fit the experimental electron density using detailed manual inspection. The N343 glycan and ACE2 (+ associated glycans) were then deleted. The equilibrated structure was then used to initiate parallel distributed MD simulations on Folding@home (Shirts and Pande, 2000, Zimmerman et al., 2020). Simulations were run with OpenMM 7.4.2 (Folding@home core22 0.0.13). Production simulations used the same Langevin integrator as the NPT equilibration described above. In total, 2995 independent MD simulations were generated on Folding@home. Conformational snapshots (frames) were stored at an interval of 1 ns/frame for subsequent analysis. The resulting final dataset contained 2995 trajectories, 1.9 ms of aggregate simulation time. Solute-only trajectories: The solute-only trajectories (with counterions) are available as MDTraj HDF5 files that contain both topology and trajectory information. A single trajectory (RUN0 CLONE0) can be downloaded using the AWS CLI:

aws s3 --no-sign-request cp s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-apo/munged/solute/17313/run0-clone0.h5 .

All HDF5 trajectories can be retrieved with

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-apo/munged/solute/17313 .

Entire dataset: The raw Folding@home dataset is made available through the AWS Open Data Registry and can be retrieved through the AWS CLI. The dataset consists of a single project (PROJ17313) and has a RUN*/CLONE*/result* directory structure. RUNs denote different equilibrated starting structures. CLONEs denote different independent replica trajectories. To retrieve raw trajectory files in gromacs XTC format for the whole dataset, you can use the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-apo/raw/PROJ17313 .

Folding@home initial files: System setup and input files can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-apo/setup-files/17313 .

Contributors: Ivy Zhang, William G. Glass, Tristan I. Croll, Aoife M. Harbison, Elisa Fadda, John D. Chodera. License: All data is freely available under the Creative Commons CC0 (“No Rights Reserved”) license.

Input and Supporting Files: ---

Trajectory: Get Trajectory (49 GB)

Represented Proteins: spike RBD

Represented Structures: 6m0j

Models: SARS-CoV-2 spike receptor-binding domain: ISOLDE refined model without N343 glycan

Trajectories of full-length SPIKE protein in the Open state. (4.2 µs )

Amaro Lab

All-atom MD simulations of full-length SPIKE protein in the Open state, protein + glycans only (not aligned). PSF and DCDs files are provided.

Input and Supporting Files: ---

Trajectory: Get Trajectory (31 GB)

Represented Proteins: spike

Represented Structures: 6VSB

Models:

Interaction between the SARS-CoV-2 spike and the α4β2 nicotinic receptor (3 replicates, 300 ns each )

A.S.F. Oliveira

University of Bristol -- Mulholland Lab

MD simulation of the complex between the Y674-R685 region of the SARS-CoV-2 spike and the extracellular domain of the human α4β2 nicotinic acetylcholine receptor. A4B2_nAChR-spike.tar.gz contains the following files. A4B2_nAChR-spike_complex.pdb A4B2_nAChR-spike_r1.tpr A4B2_nAChR-spike_r1.xtc A4B2_nAChR-spike_r2.tpr A4B2_nAChR-spike_r2.xtc A4B2_nAChR-spike_r3.tpr A4B2_nAChR-spike_r3.xtc

Input and Supporting Files:

A4B2_nAChR-spike.tar.gz

Trajectory: Get Trajectory (9 GB)

Represented Proteins: spike

Represented Structures: https://molssi-bioexcel-covid-19-structure-therapeutics-hub.s3.amazonaws.com/MulhollandGroup/nAChR-spike_interaction/A4B2_nAChR-spike_complex.pdb

Models: ---

DESRES-ANTON-11021566 10 µs simulation of of the trimeric SARS-CoV-2 spike glycoprotein, no water or ions (10 µs )

D. E. Shaw Research

DESRES

10 µs simulation trajectory of the trimeric SARS-CoV-2 spike glycoprotein with additional loop structures and glycan chains to improve the spike protein model originally released in DESRES-ANTON-[10897136,10897850]. Trajectory was initiated in the closed state (PDB entry 6VXX). The simulation used the Amber ff99SB-ILDN force field for proteins, the CHARMM TIP3P model for water, and the generalized Amber force field for glycosylated asparagine. The C- and N-peptide termini are capped with amide and acetyl groups respectively. The system was neutralized and salted with NaCl, with a final concentration of 0.15 M. The interval between frames is 1.2 ns. The simulations were conducted at 310 K in the NPT ensemble.

Input and Supporting Files:

DESRES-Trajectory_sarscov2-11021566-structure.tar.gz

DESRES-Trajectory_sarscov2-11021566.mp4

Trajectory: Get Trajectory (5.3 GB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Improved trimeric SARS-CoV-2 spike glycoprotein (closed state) in aqueous solution

Interaction between the SARS-CoV-2 spike and the αβγδ nicotinic receptor (3 replicates, 300 ns each )

A.S.F. Oliveira

University of Bristol -- Mulholland Lab

MD simulation of the complex between the Y674-R685 region of the SARS-CoV-2 spike and the extracellular domain of the αβγδ nicotinic acetylcholine receptor from Tetronarce californica (formerly Torpedo californica). ABGD_nAChR-spike.tar.gz contains the following files ABGD_nAChR-spike_complex.pdb ABGD_nAChR-spike_r1.tpr ABGD_nAChR-spike_r1.xtc ABGD_nAChR-spike_r2.tpr ABGD_nAChR-spike_r2.xtc ABGD_nAChR-spike_r3.tpr ABGD_nAChR-spike_r3.xtc

Input and Supporting Files:

ABGD_nAChR-spike.tar.gz

Trajectory: Get Trajectory (9 GB)

Represented Proteins: spike

Represented Structures: https://molssi-bioexcel-covid-19-structure-therapeutics-hub.s3.amazonaws.com/MulhollandGroup/nAChR-spike_interaction/ABGD_nAChR-spike_complex.pdb

Models: ---

Folding@home simulations of the SARS-CoV-2 spike protein (1.2 ms )

Maxwell Zimmerman

Folding@home -- Bowman lab

All-atom MD simulations of the SARS-CoV-2 spike protein, simulated using Folding@Home. The dataset comprises 3 projects, each having a RUN*/CLONE*/result* directory structure. Simulations were run using GROMACS (PROJ14217) or OpenMM (PROJ14235 and PROJ14561) and are stored as compressed binary XTC files. Each RUN represents a unique starting conformation, each CLONE is a unique MD run from the specified starting conformation, and each result is a fragment of the contiguous simulation. PROJ14217 and PROJ14253 were seeded using FAST simulations.

Topology files: The topology used in the trajectories can be downloaded directly here: PDB.

Entire dataset: The dataset is made available through the AWS Open Data Registry and can be retrieved through the AWS CLI. To retrieve raw trajectory files in gromacs XTC format for the whole dataset (7 TB), you can use the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-cryptic-pockets/SARS-CoV-2/spike/PROJ14217 .
aws s3 --no-sign-request sync s3://fah-public-data-covid19-cryptic-pockets/SARS-CoV-2/spike/PROJ14253 .
aws s3 --no-sign-request sync s3://fah-public-data-covid19-cryptic-pockets/SARS-CoV-2/spike/PROJ14561 .

Markov State Model: A polished Markov State Model (MSM), including representative cluster centers, transition probabilities, and equilibrum populations, can be downloaded using the AWS CLI. Details of how the MSM model was constructed can be found here.

aws s3 --no-sign-request sync s3://fah-public-data-covid19-cryptic-pockets/SARS-CoV-2/final_models/spike/model .

MSM cluster centers can be obtained as a gromacs XTC file from this URL: cluster centers XTC

Input files: System setup and input files can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-cryptic-pockets/SARS-CoV-2/spike/input_files .
aws s3 --no-sign-request sync s3://fah-public-data-covid19-cryptic-pockets/SARS-CoV-2/spike/PROJ14217_tpr_files .

FAST simulations: FAST simulations, which were used as seeds for Folding@Home simulations, can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-cryptic-pockets/SARS-CoV-2/FAST_simulations .

Input and Supporting Files: ---

Trajectory: Get Trajectory (6.5 TB)

Represented Proteins: spike

Represented Structures: 6VXX

Models: ---

Riken CPR TMS, TMD2_toUp trajectory (20 nanoseconds )

Takaharu Mori, Jaewoon Jung, Chigusa Kobayashi, Hisham M. Dokainish, Suyong Re, Yuji Sugita

RIKEN CPR (Cluster for Pioneering Research), TMS (Theoretical molecular science) laboratory -- TMS (Theoretical molecular science) laboratory

The data set includes a trajectory file from Targeted Molecular Dynamics (TMD) simulations of a fully glycosylated SARS-CoV-2 S-protein in solution. Water molecules and counter ions were excluded. The data includes trajectory from TMD simulation of Down to Up forms. The simulations used CHARMM36m force field for protein, and TIP3P water model. The simulations were performed using GENESIS. The coordinates were saved every 1 nanoseconds and aligned to S2 domain (Calpha atoms of residues 689-727, 854-1147).

Input and Supporting Files:

Down_pro-gly.psf

Trajectory: Get Trajectory (15MB)

Represented Proteins: spike

Represented Structures: 6vxx 6vsb

Models:

SIRAH-CoV2 initiative - RBD triple glycosylated at Asn331, 343, and 481 (10 µs )

Garay Pablo

Institut Pasteur de Montevideo -- Biomolecular Simulations Laboratory

This dataset contains the trajectory of a 10 microseconds-long coarse-grained molecular dynamics simulation of a Spike’s RBD from SARS-CoV2 glycosylated at Asn331, 343, and 481 with Man9 glycosylation trees. The initial coordinates correspond to amino acids 327 to 532 taken from the PDB structure 6XEY. Missing loops and glycosylation trees were added with CHARMM-GUI. Simulations have been performed using the SIRAH force field running with the Amber18 package at the Uruguayan National Center for Supercomputing (ClusterUY) under the conditions reported in Machado et al. JCTC 2019, adding 150 mM NaCl according to Machado & Pantano JCTC 2020. Glycan were parameterized as reported in Garay et at. 2020.

The files 6XEY-RBD-3Man9_SIRAHcg_0-4us.tar, 6XEY-RBD-3Man9_SIRAHcg_4-8us.tar, and 6XEY-RBD-3Man9_SIRAHcg_8-10us.tar, contain all the raw information required to visualize (on VMD), analyze, backmap the simulations. Step-By-Step tutorials for running, visualizing, and analyzing CG trajectories using SirahTools can be found at SIRAH website.

Additionally, the file with names ending in 6XEY-RBD-3Man9_SIRAHcg_glycoprot_10us.tar contains only the protein coordinates, while 6XEY-RBD-3Man9_SIRAHcg_glycoprot_skip10ns.tar contains one frame every 10ns.

To take a quick look at a the trajectory:

1- Untar the file 6XEY-RBD-3Man9_SIRAHcg_glycoprot_skip10ns.tar

2- Open the trajectory on VMD using the command line: vmd 6XEY-RBD-3Man9_SIRAHcg_glycoprot.prmtop 6XEY-RBD-3Man9_SIRAHcg_10us_skip10ns.ncrst 6XEY-RBD-3Man9_SIRAHcg_10us_skip10ns.nc -e sirah_vmdtk.tcl

Note that you can use normal VMD drawing methods as vdw, licorice, etc., and coloring by restype, element, name, etc.

Input and Supporting Files: ---

Trajectory: Get Trajectory (11.2 GB)

Represented Proteins: spike RBD

Represented Structures: 6XEY

Models: ---

1 microsecond trajecotry of glycosylated spike protein in closed state for pdb:6VXX embedded in viral membrane (1 µs )

Klauda lab

All atom simulation of full-glycosylated spike protein in closed state (pdb:6VXX) embedded in viral membrane. The structure was taken from Charmm-Gui at http://www.charmm-gui.org/?doc=archive&lib=covid19 where 8 models were built for the closed state. For MD simulations we used model 1-2-1 provided by Im et. al. The PSF, PDB and XTC files are uploaded

Input and Supporting Files: ---

Trajectory: Get Trajectory (12 GB)

Represented Proteins: spike

Represented Structures: 6VXX

Models:

DESRES-ANTON-11021571 10 µs simulation of of the trimeric SARS-CoV-2 spike glycoprotein, no water or ions (10 µs )

D. E. Shaw Research

DESRES

10 µs simulation trajectory of the trimeric SARS-CoV-2 spike glycoprotein with additional loop structures and glycan chains to improve the spike protein model originally released in DESRES-ANTON-[10897136,10897850]. Trajectory was initiated in a partially opened state (PDB entry 6VYB). The simulation used the Amber ff99SB-ILDN force field for proteins, the CHARMM TIP3P model for water, and the generalized Amber force field for glycosylated asparagine. The C- and N-peptide termini are capped with amide and acetyl groups respectively. The system was neutralized and salted with NaCl, with a final concentration of 0.15 M. The interval between frames is 1.2 ns. The simulations were conducted at 310 K in the NPT ensemble.

Input and Supporting Files:

DESRES-Trajectory_sarscov2-11021571-structure.tar.gz

DESRES-Trajectory_sarscov2-11021571.mp4

Trajectory: Get Trajectory (5.3 GB)

Represented Proteins: spike

Represented Structures: 6vyb

Models: Trimeric SARS-CoV-2 spike glycoprotein (open state) in aqueous solution

SIRAH-CoV2 initiative - Glycosylated RBD (10 µs )

Garay Pablo

Institut Pasteur de Montevideo -- Biomolecular Simulations Laboratory

This dataset contains the trajectories of 10 microseconds-long coarse-grained molecular dynamics simulations of SARS-CoV2 Spike´s RBD glycosylated at Asn331 and Asn343. The initial coordinates correspond to amino acids 327 to 532 taken from the PDB structure 6VSB. Missing loops and glycosylation trees were added with CHARMM-GUI.

There are two different sets of simulations corresponding to Core Complex and High Mannose. Simulations have been performed using the SIRAH force field running with the Amber18 package at the Uruguayan National Center for Supercomputing (ClusterUY) under the conditions reported in Machado et al. JCTC 2019, adding 150 mM NaCl according to Machado & Pantano JCTC 2020. Glycan were parameterized as reported in Garay et at. 2020.

The files RBD-Man9_SIRAHcg_rawdata_0-6us.tar and RBD-Man9_SIRAHcg_rawdata_6-10us.tar, contain all the raw information required to visualize (on VMD), analyze, backmap the simulations. Analogous information for Core-complex glycosylations is contained in files RBD-Core-complex_SIRAHcg_rawdata_0-6us.tar and RBD-Core-complex_SIRAHcg_rawdata_6-10us.tar.

Step-By-Step tutorials for running, visualizing, and analyzing CG trajectories using SirahTools can be found at SIRAH website.

Additionally, the file with names ending in SIRAHcg_10us_prot.tar contains only the protein coordinates, while SIRAHcg_10us_prot_skip10ns.tar contains one frame every 10ns.

To take a quick look at a the trajectory:

1- Untar the file RBD-Core-complex_SIRAHcg_10us_prot_skip10ns.tar

2- Open the trajectory on VMD using the command line: vmd RBD-Core-complex_SIRAHcg_prot.prmtop RBD-Core-complex_SIRAHcg_prot_10us_skip10ns.ncrst RBD-Core-complex_SIRAHcg_prot_10us_skip10ns.nc -e sirah_vmdtk.tcl

Note that you can use normal VMD drawing methods as vdw, licorice, etc., and coloring by restype, element, name, etc.

Input and Supporting Files: ---

Trajectory: Get Trajectory (16.4 GB)

Represented Proteins: spike RBD

Represented Structures: 6VSB

Models: ---

Cluster ensemble of 2UP like conformations (300 ns )

sugita lab

CPR

30 PDB structures of the 2Up like conformations obtained from gREST_SSCR simulations starting from 1Up conformation. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (30.8 MB)

Represented Proteins: spike

Represented Structures: 6vyb

Models: Trimeric SARS-CoV-2 spike glycoprotein (1Up state) with and without simulation box

Folding@home simulations of the SARS-CoV-2 spike RBD with P337A mutation bound to monoclonal antibody S309 (907.0 µs )

Ivy Zhang

Folding@home -- Chodera lab

All-atom MD simulations of the SARS-CoV-2 spike protein receptor binding domain (RBD) with P337A mutation bound to monoclonal antibody S309, simulated using Folding@Home. Complete details of this simulation are available here. Brief details appear below. Publication: https://doi.org/10.1038/s41586-021-03807-6 System preparation: The RBD:S309 complex was constructed from PDB ID 7JX3 (Chains A, B, and R). 7JX3 was first refined using ISOLDE to better fit the experimental electron density using detailed manual inspection. Refinement included adjusting several rotamers, flipping several peptide bonds, fixing several weakly resolved waters, and building in a missing four-residue-long loop. Though the N343 glycan N-Acetylglucosamine (NAG) was present in 7JX3, ISOLDE was used to construct a complex glycan at N343. The full glycosylation pattern was determined from Shajahan et al. and Watanabe et al. The glycan structure used for N343 (FA2G2) corresponds to the most stable conformer obtained from multi microsecond molecular dynamics (MD) simulations of cumulative sampling. The base NAG residue in FA2G2 was aligned to the corresponding NAG stub in the RBD:S309 model and any resulting clashes were refined in ISOLDE. PyMOL was used to mutate RBD’s P337 to ALA. The equilibrated structure was then used to initiate parallel distributed MD simulations on Folding@home (Shirts and Pande, 2000, Zimmerman et al., 2020). Simulations were run with OpenMM 7.4.2 (Folding@home core22 0.0.13). Production simulations used the same Langevin integrator as the NPT equilibration described above. In total, 4998 independent MD simulations were generated on Folding@home. Conformational snapshots (frames) were stored at an interval of 1 ns/frame for subsequent analysis. The resulting final dataset contained 4998 trajectories, 907.0 µs of aggregate simulation time. Solute-only trajectories: The solute-only trajectories (with counterions) are available as MDTraj HDF5 files that contain both topology and trajectory information. A single trajectory (RUN0 CLONE0) can be downloaded using the AWS CLI:

aws s3 --no-sign-request cp s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/munged/solute/17342/run0-clone0.h5 .

All HDF5 trajectories can be retrieved with

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/munged/solute/17342 .

Entire dataset: The raw Folding@home dataset is made available through the AWS Open Data Registry and can be retrieved through the AWS CLI. The dataset consists of a single project (PROJ17342) and has a RUN*/CLONE*/result* directory structure. RUNs denote different equilibrated starting structures. CLONEs denote different independent replica trajectories. To retrieve raw trajectory files in gromacs XTC format for the whole dataset, you can use the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/raw-data/PROJ17342 .

Folding@home initial files: System setup and input files can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/setup-files/17342 .

Contributors: Ivy Zhang, William G. Glass, Tristan I. Croll, Aoife M. Harbison, Elisa Fadda, John D. Chodera. License: All data is freely available under the Creative Commons CC0 (“No Rights Reserved”) license.

Input and Supporting Files: ---

Trajectory: Get Trajectory (89 GB)

Represented Proteins: spike RBD

Represented Structures: 7jx3

Models: SARS-CoV-2 spike receptor-binding domain bound with S309: ISOLDE refined model with N343 glycan and P337A mutation

Cluster ensemble of 1UP like conformation (500 ns )

sugita lab

CPR

30 PDB structures of the 1Up like cluster obtained from gREST_SSCR simulations starting from Down conformation. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (30.8 MB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Trimeric SARS-CoV-2 spike glycoprotein (Down state) with and without simulation box

Folding@home simulations of the SARS-CoV-2 spike RBD bound to monoclonal antibody S309 (1.1 ms )

Ivy Zhang

Folding@home -- Chodera lab

All-atom MD simulations of the SARS-CoV-2 spike protein receptor binding domain (RBD) bound to monoclonal antibody S309, simulated using Folding@Home. Complete details of this simulation are available here. Brief details appear below. Publication: https://doi.org/10.1038/s41586-021-03807-6 System preparation: The RBD:S309 complex was constructed from PDB ID 7JX3 (Chains A, B, and R). 7JX3 was first refined using ISOLDE to better fit the experimental electron density using detailed manual inspection. Refinement included adjusting several rotamers, flipping several peptide bonds, fixing several weakly resolved waters, and building in a missing four-residue-long loop. Though the N343 glycan N-Acetylglucosamine (NAG) was present in 7JX3, ISOLDE was used to construct a complex glycan at N343. The full glycosylation pattern was determined from Shajahan et al. and Watanabe et al. The glycan structure used for N343 (FA2G2) corresponds to the most stable conformer obtained from multi microsecond molecular dynamics (MD) simulations of cumulative sampling. The base NAG residue in FA2G2 was aligned to the corresponding NAG stub in the RBD:S309 model and any resulting clashes were refined in ISOLDE. The equilibrated structure was then used to initiate parallel distributed MD simulations on Folding@home (Shirts and Pande, 2000, Zimmerman et al., 2020). Simulations were run with OpenMM 7.4.2 (Folding@home core22 0.0.13). Production simulations used the same Langevin integrator as the NpT equilibration described above. In total, 5000 independent MD simulations were generated on Folding@home. Conformational snapshots (frames) were stored at an interval of 1 ns/frame for subsequent analysis. The resulting final dataset contained 5000 trajectories, 1.1 ms of aggregate simulation time. Solute-only trajectories: The solute-only trajectories (with counterions) are available as MDTraj HDF5 files that contain both topology and trajectory information. A single trajectory (RUN0 CLONE0) (~42 MB) can be downloaded using the AWS CLI:

aws s3 --no-sign-request cp s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/munged/solute/17341/run0-clone0.h5 .

All HDF5 trajectories can be retrieved with

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/munged/solute/17341 .

Entire dataset: The raw Folding@home dataset is made available through the AWS Open Data Registry and can be retrieved through the AWS CLI. The dataset consists of a single project (PROJ17341) and has a RUN*/CLONE*/result* directory structure. RUNs denote different equilibrated starting structures. CLONEs denote different independent replica trajectories. To retrieve raw trajectory files in gromacs XTC format for the whole dataset, you can use the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/raw-data/PROJ17341 .

Folding@home initial files: System setup and input files can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-RBD-antibody/setup-files/17341 .

Contributors: Ivy Zhang, William G. Glass, Tristan I. Croll, Aoife M. Harbison, Elisa Fadda, John D. Chodera. License: All data is freely available under the Creative Commons CC0 (“No Rights Reserved”) license.

Input and Supporting Files: ---

Trajectory: Get Trajectory (102 GB)

Represented Proteins: spike RBD

Represented Structures: 7jx3

Models: SARS-CoV-2 spike receptor-binding domain bound with S309: ISOLDE refined model with N343 glycan

DESRES-ANTON-10897850 10 µs simulation of of the trimeric SARS-CoV-2 spike glycoprotein, no water or ions (10 µs )

D. E. Shaw Research

DESRES

A 10 µs simulation of the trimeric SARS-CoV-2 spike glycoprotein. System was initiated in a partially opened state (PDB entry 6VYB) which exhibited a high degree of conformational heterogeneity. In particular, the partially detached receptor binding domain sampled a variety of orientations, and further detached from the S2 fusion machinery. The simulation used the Amber ff99SB-ILDN force field for proteins, the CHARMM TIP3P model for water, and the generalized Amber force field for glycosylated asparagine. The C- and N-peptide termini, including those exposed due to missing loops in the published structural models, are capped with amide and acetyl groups respectively. The system was neutralized and salted with NaCl, with a final concentration of 0.15 M. The total number of atoms in the system was 715439 for the closed state. The interval between frames is 1.2 ns. The simulations were conducted at 310 K in the NPT ensemble. We have released new versions of these simulations with enhancements to the spike protein model in [DESRES-ANTON-11021566,11021571] (https://www.deshawresearch.com/downloads/download_trajectory_sarscov2.cgi/#DESRES-ANTON-11021566), since the one used in this simulation is incomplete in some of the disordered loop regions (i.e., resid 455 to 461, resid 469 to 488) and in glycan chains.

Input and Supporting Files:

DESRES-Trajectory_sarscov2-10897850-structure.tar.gz

DESRES-Trajectory_sarscov2-10897850.mp4

Trajectory: Get Trajectory (4.1 GB)

Represented Proteins: spike

Represented Structures: 6vyb

Models: Trimeric SARS-CoV-2 spike glycoprotein (open state) in aqueous solution

Riken CPR TMS, MD1_Down trajectory (1 microseconds )

Takaharu Mori, Jaewoon Jung, Chigusa Kobayashi, Hisham M. Dokainish, Suyong Re, Yuji Sugita

RIKEN CPR (Cluster for Pioneering Research), TMS (Theoretical molecular science) laboratory -- TMS (Theoretical molecular science) laboratory

The data set includes a trajectory file from Molecular Dynamics (MD) of a fully glycosylated SARS-CoV-2 S-protein in solution. Water molecules and counter ions were excluded. The starting structure is an inactive Down taken from CHARMM-GUI COVID-19 Archive (http://www.charmm-gui.org/docs/archive/covid19). We replaced counter ions K+ in the original model with Na+. The simulation used CHARMM36m force field for protein, and TIP3P water model. The simulation was performed using GENESIS. The coordinates were saved every 1 nanoseconds and aligned to S2 domain (Calpha atoms of residues 689-727, 854-1147).

Input and Supporting Files:

Down_pro-gly.psf

Trajectory: Get Trajectory (742MB)

Represented Proteins: spike

Represented Structures: 6vxx

Models:

Simulations of SARS-CoV and SARS-CoV-2 RBD with ACE2 (2 µs )

Gumbart lab

Two-microsecond trajectories of the receptor-binding domains from SARS-CoV and SARS-CoV-2 spike protein bound to the human receptor, ACE2 (two replicas each). Simulation systems were constructed with VMD, equilibrated initially with NAMD, and then run for 2 µs each with Amber16. Simulations used a 4-fs timestep enabled by hydrogen-mass repartitioning (HMR).

Input and Supporting Files: ---

Trajectory: Get Trajectory (5.3 GB)

Represented Proteins: spike RBD ACE2

Represented Structures: 2AJF 6M17

Models: ---

Continuous trajectories of glycosylated SPIKE opening. (175 ns )

Amaro Lab and Chong Lab

All-atom MD trajectories from weighted ensemble simulations of glycosylated SPIKE protein, protein + glycans only. PSF, prmtop, DCDs, and WESTPA input files are provided. Starting structure based on model of the full-length spike in the closed state developed by the Amaro lab, which is modeled from 6VXX. Only the head region of the Spike was included in simulations from residues 16-1140.

Input and Supporting Files: ---

Trajectory: Get Trajectory (1.35 GB)

Represented Proteins: spike

Represented Structures: 6VXX

Models:

Trajectories of full-length SPIKE protein in the Closed state. (1.7 µs )

Amaro Lab

All-atom MD simulations of full-length SPIKE protein in the Closed state, protein + glycans only (not aligned). PSF and DCDs files are provided.

Input and Supporting Files: ---

Trajectory: Get Trajectory (13 GB)

Represented Proteins: spike

Represented Structures: 6VXX

Models:

Cluster ensemble of Intermediate 3a (500 ns )

sugita lab

CPR

30 PDB structures of the intermediate (I3a) cluster obtained from gREST_SSCR simulations starting from Down conformation. Water molecules and Ions are removed from these PDB structures.

Input and Supporting Files: ---

Trajectory: Get Trajectory (30.8 MB)

Represented Proteins: spike

Represented Structures: 6vxx

Models: Trimeric SARS-CoV-2 spike glycoprotein (Down state) with and without simulation box

MD simulations of trimeric SARS-Cov2 spike protein ectodomain in explicit solvent. Data were collected for apo, linoleic acid bound and other putative ligands (3x200 ns in each case) (24 x 200 ns trajectories (solvent removed) )

Deborah K Shoemark

University of Bristol, UK -- BrisSynBio and Mulholland

The CryoEM stuctures of the apo and linoleic acid bound SARS-Cov2 spike protein trimer (residues 15/25 to 1139) were used to build complete atomistic models. Other putative ligands, including cholesterol and vitamins, retinoids and steroids identified by docking with BUDE, were simulated in both open and closed states. The closed and open structures have 42 and 43 disulfide bonds respectively. Simulations were performed with GROMACS 2019.x. the file Spike_MD_simulations.tgz contains:

• Spike_MD_simulations/
• Spike_MD_simulations/WT_closed-SARS2-spike_apo/
• Spike_MD_simulations/WT_closed-SARS2-spike_apo/01_WT_closed_apo_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_apo/02_WT_closed_apo_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_apo/03_WT_closed_apo_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_apo/01_WT_closed_apo_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_apo/02_WT_closed_apo_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_apo/03_WT_closed_apo_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_apo/README
• Spike_MD_simulations/WT_closed-SARS2-spike_cholesterol/
• Spike_MD_simulations/WT_closed-SARS2-spike_cholesterol/01_WT_closed-OK_CLR_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_cholesterol/02_WT_closed-OK_CLR_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_cholesterol/03_WT_closed-OK_CLR_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_cholesterol/01_WT_closed-OK_CLR_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_cholesterol/02_WT_closed-OK_CLR_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_cholesterol/03_WT_closed-OK_CLR_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_cholesterol/README
• Spike_MD_simulations/WT_closed-SARS2-spike_dexamethasone/
• Spike_MD_simulations/WT_closed-SARS2-spike_dexamethasone/01_clean_WT_closed_dexys_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_dexamethasone/02_clean_WT_closed_dexys_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_dexamethasone/03_clean_WT_closed_dexys_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_dexamethasone/README
• Spike_MD_simulations/WT_closed-SARS2-spike_dexamethasone/01_clean_WT_closed_dexys_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_dexamethasone/02_clean_WT_closed_dexys_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_dexamethasone/03_clean_WT_closed_dexys_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_LA/
• Spike_MD_simulations/WT_closed-SARS2-spike_LA/01_WT_closed_LA_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_LA/02_WT_closed_LA_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_LA/03_WT_closed_LA_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_closed-SARS2-spike_LA/01_WT_closed_LA_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_LA/02_WT_closed_LA_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_LA/03_WT_closed_LA_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_closed-SARS2-spike_LA/README
• Spike_MD_simulations/WT_open-SARS2-spike_apo/
• Spike_MD_simulations/WT_open-SARS2-spike_apo/01_WT-OK_open_apo_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_apo/02_WT-OK_open_apo_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_apo/03_WT-OK_open_apo_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_apo/README
• Spike_MD_simulations/WT_open-SARS2-spike_apo/01_WT-OK_open_apo_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_apo/02_WT-OK_open_apo_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_apo/03_WT-OK_open_apo_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_cholesterol/
• Spike_MD_simulations/WT_open-SARS2-spike_cholesterol/01_WT-open-OK_CLR_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_cholesterol/01_WT-open-OK_CLR_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_cholesterol/02_WT-open-OK_CLR_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_cholesterol/02_WT-open-OK_CLR_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_cholesterol/03_WT-open-OK_CLR_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_cholesterol/03_WT-open-OK_CLR_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_cholesterol/README
• Spike_MD_simulations/WT_open-SARS2-spike_dexamethasone/
• Spike_MD_simulations/WT_open-SARS2-spike_dexamethasone/01_clean_WT_open_dexys_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_dexamethasone/03_clean_WT_open_dexys_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_dexamethasone/01_clean_WT_open_dexys_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_dexamethasone/03_clean_WT_open_dexys_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_dexamethasone/02_clean_WT_open_dexys_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_dexamethasone/02_clean_WT_open_dexys_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_dexamethasone/README
• Spike_MD_simulations/WT_open-SARS2-spike_LA/
• Spike_MD_simulations/WT_open-SARS2-spike_LA/01_WT-OK_open_LAs_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_LA/02_WT-OK_open_LAs_200_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_LA/03_WT-OK_open_LAs_200ns_mol_noj_fit.xtc
• Spike_MD_simulations/WT_open-SARS2-spike_LA/01_WT-OK_open_LAs_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_LA/02_WT-OK_open_LAs_200_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_LA/03_WT-OK_open_LAs_200ns_mol_noj_fit.pdb
• Spike_MD_simulations/WT_open-SARS2-spike_LA/README
• Spike_MD_simulations/README

Input and Supporting Files:

Spike_MD_simulations.tgz

Trajectory: Get Trajectory (9 GB)

Represented Proteins: spike

Represented Structures: 6ZB5

Models: ---

Folding@home simulations of the SARS-CoV-2 spike RBD with N501Y mutation bound to human ACE2 (953.7 µs )

Ivy Zhang

Folding@home -- Chodera lab

All-atom MD simulations of the SARS-CoV-2 spike protein receptor binding domain (RBD) with N501Y mutation bound to human angiotensin converting enzyme-related carboypeptidase (ACE2), simulated using Folding@Home. Complete details of this simulation are available here. Brief details appear below. Publication: https://doi.org/10.1016/j.cell.2021.01.037 System preparation: The RBD:ACE2 complex was constructed from individual RBD (PDB: 6m0j, Chain E) and ACE2 (PDB: 1r42, Chain A) monomers aligned to the full RBD:ACE2 structure (PDB: 6m0j. These structural models were further refined by Tristan Croll using ISOLDE (Croll, 2018) and deposited in the Coronavirus Structural Taskforce (CST) database (Croll et al., 2020) to produce refined 6m0j and refined 1r42 models. The RBD N501 was mutated to TYR using PyMOL 2.3.2. The resulting RBD and ACE2 monomers were then aligned in PyMOL 2.3.2 to the CST 6m0j structure to create an initial RBD:ACE2 complex. Full glycosylation patterns for ACE2 and RBD glycans were determined from Shajahan et al. For the constructed RBD:ACE2 complex, these included sites: N53, N90, N103, N322, N432, N546, and N690 on ACE2 and N343 on the RBD. Base NAG residues of each glycan structure (FA2, FA26G1, FA2, FA2, FA2G2, A2, FA2, FA2G2, respectively) were acquired from Elisa Fadda. Each glycan was then aligned to the corresponding NAG stub in the RBD:ACE2 model in and any resulting clashes were refined in ISOLDE. Full details of the glycosylation patterns / structures used and full workflow are available here. Folding@home simulation: The equilibrated structure was then used to initiate parallel distributed MD simulations on Folding@home (Shirts and Pande, 2000, Zimmerman et al., 2020). Simulations were run with OpenMM 7.4.2 (Folding@home core22 0.0.13). Production simulations used the same Langevin integrator as the NpT equilibration described above. In total, 5000 independent MD simulations were generated on Folding@home. Conformational snapshots (frames) were stored at an interval of 1 ns/frame for subsequent analysis. The resulting final dataset contained 5000 trajectories and 953.7 µs of aggregate simulation time. Solute-only trajectories: The solute-only trajectories (with counterions) are available as MDTraj HDF5 files that contain both topology and trajectory information. A single trajectory of the WT RBD (RUN3) can be downloaded using the AWS CLI:

aws s3 --no-sign-request cp s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-ACE2-RBD/munged/solute/17344/run0-clone0.h5 .

All HDF5 trajectories can be retrieved with

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-ACE2-RBD/munged/solute/17344 .

Entire dataset: The raw Folding@home dataset is made available through the AWS Open Data Registry and can be retrieved through the AWS CLI. The dataset consists of a single project (PROJ17344) and has a RUN*/CLONE*/result* directory structure. RUNs denote different equilibrated starting structures. CLONEs denote different independent replica trajectories. To retrieve raw trajectory files in gromacs XTC format for the whole dataset, you can use the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-ACE2-RBD/raw-data/PROJ17344 .

Folding@home initial files: System setup and input files can be downloaded using the AWS CLI:

aws s3 --no-sign-request sync s3://fah-public-data-covid19-antibodies/vir-collaboration/SARS-CoV-2-ACE2-RBD/setup/17344 .

Contributors: Ivy Zhang, William G. Glass, Tristan I. Croll, Aoife M. Harbison, Elisa Fadda, John D. Chodera. License: All data is freely available under the Creative Commons CC0 (“No Rights Reserved”) license.

Input and Supporting Files: ---

Trajectory: Get Trajectory (132 GB)

Represented Proteins: spike RBD ACE2

Represented Structures: 6m0j 1R42

Models: SARS-CoV-2 spike receptor-binding domain: ISOLDE refined model with N343 glycan and N501Y mutation Native Human Angiotensin Converting Enzyme-Related Carboxypeptidase (ACE2): ISOLDE refined model with glycans

Riken CPR TMS, TMD2_toDown trajectory (20 nanoseconds )