# Markov state models of Folding@home COVID19 simulations
In response to the COVID19 pandemic, [Folding@home](https://foldingathome.org/) users ran simulations of strategic value for understanding the molecular mechanisms of SARS-CoV-2 and its pathology. For detailed information on this dataset and core highlights, please view our manuscript, ["SARS-CoV-2 Simulations Go Exascale to Capture Spike Opening and Reveal Cryptic Pockets Across the Proteome"](https://www.biorxiv.org/content/10.1101/2020.06.27.175430v3).
## Data Format
Systems are listed by organism followed by protein. Within each entry there is a `README.dat` for metadata on Markov State Model construction.
### Markov State Model
The Markov State Model is contained within the `model` directory, which contains:
- `tprobs.npy` - Transition probability matrix in numpy binary format
- `tcounts.npy` - Transition counts matrix in numpy binary format
- `prot_masses.pdb` - Protein topology file in PDB format
- `populations.npy` - Equilibrium populations for defined states
- `full_centers.xtc` - Representative cluster centers stored as a GROMACS compressed XTC file
### cryptic pockets
Each system was assesed for cryptic pockets with analysis listed in the `cryptic_pockets` directory.
The residues that comprise a cryptic pocket are provided in the text file `pocketXX_resis.dat`, where XX denotes the pocket numbeer. Additionally, cluster centers are ranked by size of cryptic pocket opening and sorted from highest to lowest. `pocketXX_rankings.dat` contains this sorted list, where the first entry is the cluster center with the largest pocket opening, and the last entry is the cluster centers with the smallest pocket opening.
## loading models and data
Cluster centers can be visualized using [VMD](https://www.ks.uiuc.edu/Research/vmd/):
```
> vmd prot_masses.pdb full_centers.xtc
```
Additionally, structural analysis can be performed in python using the package [MDTraj](http://mdtraj.org/1.9.3/)
```
import mdtraj as md
# load trajectory
trj = md.load("full_centers.xtc", top="prot_masses.pdb")
# compute distances between atom indices 0 and 1
distances = md.compute_distances(trj, atom_pairs=[[0, 1]])
# save each cluster center as a pdb
for state_num in range(trj.n_frames):
trj[state_num].save_pdb("state%d.pdb" % state_num)
```
