深圳市计算化学与药物设计重点实验室
Lab of Computational Chemistry and Drug Design

PBSP (Phosphate Binding Site Predictor) is a stand-alone command line program that predicts phosphate binding sites from protein structures. The approach leverages a modified energy-based ligand binding sites identification method and combines it with reverse focused docking in which a phosphate probe is used to improve the accuracy and selectivity of predictions. Here, we developed two PBSP models, PBSP_R by rigid docking and PBSP_F by flexible docking.

Download PBSP and Benchmark Datasets

(1) PBSP_R (run on linux 64 bit)

(2) PBSP_F (run on linux 64 bit)

(3) Bound dataset

a non-redundant bound dataset of 97 phosphate binding proteins containing 106 phosphate binding sites (proteins < 50% sequence identity).

(4) Unbound dataset

a non-redundant unbound dataset of 63 phosphate binding proteins containing 67 phosphate binding sites (proteins < 50% sequence identity).

Installation

PBSP relies on and contains third-party softwares called AutoDockFR and AutoSite. The PBSP executable should be place in the same folder as the modified ADFRsuite. A  linux system(64 bit) are required to run the scripts that are provided to automate the calculations. The suite does not require any special installation procedure. The archive downloaded from the web must be uncompressed and can then be placed anywhere.

Usage

The PBSP  executable provided in the root directory of the suite allows the user to carry out phosphate binding site identification using focused reverse docking. An example of the most basic use:

PBSP_R --input=protein.pdb --probe=CH3_PO4.pdb --outdir=/lustre/home/zclu/

PBSP_R -i protein.pdb -p CH3_PO4.pdb -o /lustre/home/zclu/

CH3PO42- is a phosphate probe which is docked to the potential phosphate binding pocket to generate phosphate binding modes.

Tips

1) PBSP_R is a executable which are generated by from the source Python script 'PBSP_R.py' by PyInstaller.
2) PyInstaller is a program that converts Python programs into stand-alone executables, under Windows, Linux, and Mac OS X. PyInstaller can bundles a Python application and all its dependencies into a single package. The user can run the packaged app without installing a Python interpreter or any modules.
3) If you are familiar with Python programming languages, you can use and modify the the script 'PBSP_R.py'. The 'PBSP_R.py' script is written in Python 3, and is placed in the script floder.

Outputs

(1) Predicted phosphate binding sites
Output file: 1d4w_C_281_predicted_phosphate_binding_sites.txt

Rank     Pocket     Predicted Phosphate Binding Sites     Predicted Binding Energy
1   1d4w_C_281_pocket000     ARG_55,GLU_35,SER_36,SER_34,CYS_42,ARG_32     -5.7
2   1d4w_C_281_pocket001     LYS_81,LYS_79     -5.5
3   1d4w_C_281_pocket013     ARG_75,LYS_89,GLN_92     -3.7
4   1d4w_C_281_pocket008     LYS_74,THR_68     -3.5
5   1d4w_C_281_pocket003     SER_65,LYS_74,GLU_67,SER_57     -3.4
6   1d4w_C_281_pocket009     SER_65,LYS_74,SER_57     -3.4
7   1d4w_C_281_pocket002     GLY_93,ASP_91,PRO_97,ILE_94,GLN_92     -3.3
8   1d4w_C_281_pocket021     ASN_82,ARG_78     -3.2
9   1d4w_C_281_pocket022     TYR_41,MET_1     -3.1
10   1d4w_C_281_pocket004   ASP_2,LYS_104,GLY_9     -2.9
11   1d4w_C_281_pocket010   GLN_58,SER_57     -1.7
12   1d4w_C_281_pocket011   GLU_14,ARG_13,GLU_35     -1.7
13   1d4w_C_281_pocket006   GLY_39,PRO_38,VAL_40,SER_34,ASP_33     -1.6
14   1d4w_C_281_pocket012   TYR_29,GLN_88     -1.6
15   1d4w_C_281_pocket015   ARG_32,ARG_13,SER_12,GLU_35     -1.1
16   1d4w_C_281_pocket024   GLN_99,TYR_100     -1.1
17   1d4w_C_281_pocket023   GLU_103,VAL_102     -1.0

(2) Predicted binding modes
Output file: 1d4w_C_281_pocket000_complex.pdb

Citation

PBSP relies on and contains third-party softwares called AutoDockFR and AutoSite. if you used PBSP in your work, please cite:

1) Z.C. Lu, F. Jiang and Y.D. Wu, Phosphate binding sites prediction in phosphorylation-dependent protein-protein interactions, in revision

2) P.A. Ravindranath, S. Forli, D.S. Goodsell, A.J. Olson and M.F. Sanner, AutoDockFR: Advances in Protein-Ligand Docking with Explicitly Specified Binding Site Flexibility, PLoS Comput Biol (2015). DOI:10.1371/journal.pcbi.1004586

3) P.A. Ravindranath and M.F. Sanner, AutoSite: an automated approach for pseudo- ligands prediction - From ligand binding sites identification to predicting key ligand atoms, Bioinformatics (2016). DOI:10.1093/bioinformatics/btw367

Please contact Zheng-Chang Lu (luzc@pku.edu.cn) if you have an idea for a project proposal or any questions.