https://doi.org/10.5061/dryad.4b8gthtmf

The following is an explanation of the overall dataset submitted to this repository:

1. Stripped_Parm.tar.gz (2.69 MB): parameter files used for the simulations. Files are in AMBER22 format.
2. Dtraj_Weights.tar.gz (15.57 MB): Trajectory weights obtained after MSM construction. These were then used to reweigh the ensembles shown in Figures 3, 4, 5, 6.
   • To construct the MSM, a set of 57 distances was chosen (described in the manuscript). time-lagged Independent Component Analysis (tICA) was performed on these set of distances, which uses a linear combination of the 57 distances to reduce the dimensions, where the first dimension represented the slowest kinetic process (activation) observed in the system. Then, k-means clustering was performed on the tIC dimensions and an MSM was constructed by optimizing the VAMP2 score after choosing the optimal number of clusters, number of tICs and MSM lagtime.
   • Code used to construct the MSM and get the weights is available here.
3. PDB_files.tar.gz (10.35 MB): The starting points used for the simulations.
4. PDB_Distances_57.tar.gz (1.66 KB): The 57 distances used to construct the MSM were calculated for the starting points to refer to them in Figures 3, 4, 5.
5. WGM_plot_CRD.tar.gz (1.41 GB): .npy files used to plot Fig. 3B in the manuscript. These files contain data of collective variables that were used to explain the function of the WGM motif in SMO activation, when CYC is bound to the CRD of SMO.
   • The collective variables were calculated using the script available here. The code used to plot these .npy files is available here.
6. WGM_plot_Daul.tar.gz (2.28 GB): .npy files used to plot Fig. 3D in the manuscript. These files contain data of collective variables that were used to explain the function of the WGM motif in SMO activation, when CYC is bound to both sites - the CRD and the TMD of SMO.
   • The collective variables were calculated using the script available here. The code used to plot these .npy files is available here.
7. WGM_plot_TMD.tar.gz (2.26 GB): .npy files used to plot Fig. 3F in the manuscript. These files contain data of collective variables that were used to explain the function of the WGM motif in SMO activation, when CYC is bound to the TMD of SMO.
   • The collective variables were calculated using the script available here. The code used to plot these .npy files is available here.
8. DREnpy_CRD.tar.gz (98.79 MB): .npy files used to plot Fig. 4B in the manuscript. These files contain data of collective variables that were used to explain the function of the DRE lock in SMO activation, when CYC is bound to the CRD of SMO.
   • The collective variables were calculated using the script available here. The code used to plot these .npy files is available here.
9. DREnpy_Dual.tar.gz (160.05 MB): .npy files used to plot Fig. 4D in the manuscript. These files contain data of collective variables that were used to explain the function of the DRE lock in SMO activation, when CYC is bound to both sites - the CRD and the TMD of SMO.
   • The collective variables were calculated using the script available here. The code used to plot these .npy files is available here.
10. DREnpy_TMD.tar.gz (157.60 MB): .npy files used to plot Fig. 4F in the manuscript. These files contain data of collective variables that were used to explain the function of the DRE lock in SMO activation, when CYC is bound to the TMD of SMO.
    • The collective variables were calculated using the script available here. The code used to plot these .npy files is available here.
11. extracted_frames.tar.gz (9.60 GB): .pdb files representing sampled from the MSM for each ensemble. These PDBs are representative of the Active, Inactive and Intermediate states observed during the simulations for each ensemble. These were then used to compute the tunnel profiles shown in Fig. 6 of the manuscript.
12. Tunnel_Files.tar.gz (104.95 MB): .npy files containing the tunnel profiles shown in Fig. 6 of the manuscript. These files contain data that enables us to differentiate the tunnel radii when cyclopamine is bound to different binding sites in SMO.
    • The collective variables were calculated using the script available here. The code used to plot these .npy files is available here.
13. Random_Forest_Files.tar.gz (930.36 MB): .npy files used to train the Random Forest Classifier used to distinguish between ensembles. Methodology explained in the manuscript.
    • The model training code is available here.
14. Angle_Files.tar.gz (1.13 MB) : List of .npy files used to plot the kernel density plot for Figure 7 of the paper. These files contain data of collective variables that distinguish the three ensembles described in the paper.
    • The code to calculate these metrics is available here. The code used to plot these .npy files is available here.
15. fig2c_data.pkl (0.8kb) and alchemical_plot_data_fig2b.pkl (0.4kb) : List of files used to make Fig 2b and Fig 2c for the paper. These files contain data of the alchemical free energies and the equilibrium populations.
    • The code to calculate these metrics is available here. The code used to plot these .npy files is available here.
16. alchemical_data.tar.gz: (1.95 GB): This contains the 5 datasets used to compute the alchemical free energies presented in Fig. 2b of the study. The tarball contains 5 folders, each containing the dHdl.xvg files used to compute the alchemical free energies. 
    • The code to compute the free energies from these files is available here.

Links to other publicly accessible locations of the data (Box link, total size > 800 GB):

• https://uofi.box.com/s/4g3xmumfmesb68y7tb0fn8wvhvycylr

Code/Software

Code used to analyze the trajectories and construct the Markov Model is openly available at https://github.com/ShuklaGroup/SMO_CYC/