debbydanwang/IFPscore

IFP-based scoring functions

IFPscore

IFP-based scoring functions

1. Requirements IFPscore currently supports a Linux system and Python 3.6, and requires main dependency packages as follows.

   • deemchem (https://github.com/deepchem/deepchem)
   • rdkit (https://www.rdkit.org/)
   • numpy (https://numpy.org/)
   • sklearn (https://scikit-learn.org/stable/)
   • multiprocessing (https://docs.python.org/3/library/multiprocessing.html)
   • Bio (https://biopython.org/)
   • scipy (https://www.scipy.org/)

2. Data downloading and preprocessing

   1. Downloading:

      • Model construction 'PDBbind_refined' data folder: PDBbind refined set (http://www.pdbbind.org.cn/)
      • validation
        • 'PDBbind_core' data folder: PDBbind core set (http://www.pdbbind.org.cn/)
        • 'csarhiqS1' data folder: CSAR-HiQ sets 1 (http://www.csardock.org/)
        • 'csarhiqS2' data folder: CSAR-HiQ sets 2 (http://www.csardock.org/)
        • 'csarhiqS3' data folder: CSAR-HiQ sets 3 (http://www.csardock.org/)

   2. Preprocessing:

      • PDBbind refined/core sets: save the ligand files as PDB files (e.g. using software like UCSF Chimera)
      • CSAR-HiQ sets: save the protein and ligand in each complex as PDB files (e.g. using software like UCSF Chimera), and name these files as those in PDBbind refined/core sets (e.g. 1ax1_protein.pdb, 1ax1_ligand.pdb in '1ax1' folder)

   3. Put these folders together:

      • Create a folder (e.g. 'Score') and put all these data folders (e.g. 'PDBbind_refined', 'PDBbind_core', 'csarhiqS1', 'csarhiqS2', 'csarhiqS3') and the index folder ('indexes' folder in this repository) in 'Score'

3. Example codes are provided in the 'Examples' folder in this repository

   1. PrtCmmIFPScore - Constructing a PrtCmm IFP Score on PDBbind refined set (excluding the validation sets) and validating it on the four validation sets (PDBbind core set and CSAR-HiQ sets) using Pearson's correlation and RMSE
   2. RFprtcmmScore - Constructing an RF-SCORE (PrtCmm version) on PDBbind refined set (excluding the validation sets) and validating it on the four validation sets (PDBbind core set and CSAR-HiQ sets) using Pearson's correlation and RMSE