DeepBindBC is a deep learning based binary classifier for identifying native-like protein-ligand complexes

Extract pocket by known ligand(optional)

(Needed when user is not sure how to prepare protein pocket information)



Input example

python script to extract pocket by known ligand in correct binding position
python script to extract pocket by given pocket position

Submit a job

(One receptor against multiple ligand with known pocket, the protein name should be ????.pdb,the pocket name should be ????_pocket.pdb.)


Input example

(The output summary file is summary_all.txt, column 1 is ligand name with its conformation id, column 2 is DeepBindBC's predicted possiblity value, column 3 is vina docking score)

Submit a job

Fast finding out the type of ligands that prefer to binding with the given protein with known pocket (One receptor against 100 representative ligands from cluster of a FDA approved drug dataset, the protein name should be ????.pdb,the pocket name should be ????_pocket.pdb, and they should put in a folder named receptor and zipped, exactly as the following example)


Input example

(The output summary file is summary_all.txt, column 1 is ligand name with its conformation id, column 2 is DeepBindBC's predicted possiblity value, column 3 is vina docking score)

References

    A DEEP LEARNING METHOD AS A CORE COMPONENT OF VIRTUAL SCEENING PIPELINE WITH ITS APPLICATION IN PANCREATIC ALPHA-AMYLASE, Haiping Zhang, Tingting Zhang, Konda Mani Saravanan, Linbu Liao, Hao Wu, Haishan Zhang, Huiling Zhang, Yi Pan, Xuli Wu*, Yanjie Wei* (submitted)