RegDock(1)                   User Commands                   RegDock(1)

NAME

     RegDock v0.1

DOWNLOAD

RegDock v0.1
An example consists of 2 PDB files.

SYNOPSIS

Currently, RegDock software consists of several binary files (.exe), and several batch files (.bat).

The executional file list are as follows:

!1_pdb2ncd1.bat
call pdb2xyz.exe, msms.exe, and vert2ncd.exe to convert protein file (.pdb) to point cloud file (.ncd).
The msms.exe is not in our distribution, and could be downloaded at MSMS_HOME.

!2_cut.bat
call CloudCut.exe to split point cloud of molecular surface into small pieces.

!3_docking.bat
call RegDock.exe to generate coarse matching result. (global matching) This file uses some matlab features, thus user have to install matlab before using.

!4_cluster.bat
call tranprocess.exe to merge and cluster the coarse matching results.

!5_local.bat
call pdb2xyz.exe, msms.exe, and vert2ncd.exe to generate sparse point cloud for local matching.
Then, call cloudreverse.exe and RegDock.exe to refine the matching result. (local matching)

!6_scoring.bat
call RegDock.exe to calculate score for the matching results.

DESCRIPTION

RegDock computes, for two given moleculars receptor A and ligand B, the possible docking results, and the geometric complementary score of each result. The docking results are given by the rigid body motion of B from its initial location.

Components

For a tipical usage of RegDock, user should configure the .bat files, and run them one by one. Some of the steps could be altered due to the user's needs.

Read PDB step (!1_pdb2ncd1.bat)

In this step, the two protein files are read in, and two point cloud files are generated. Input: The two input file should be in PDB format, and has file naming as A.pdb and B.pdb. Configure: Replace '1PPE_E' with 'A', and '1PPE_I' with 'B' in file '!1_pdb2ncd1.bat'. Make sure that the pdb files and the executable files are in the same directory. Output: The output of execution is two point cloud files naming as A.ncd and B.ncd.

Cut NCD step (!2_cut.bat)

In this step, the ligand point cloud is cut into smaller pieces for surface matching, the smaller pieces could be intersection to each other. Input: The input file should be in .ncd format. Configure: Replace '1PPE_I.ncd' with the input file name in file '!2_cut.bat'. The second '1PPE_I.ncd' in file '!2_cut.bat' is the common suffix of the output files. The 3rd parameter controls the size of the result point clouds, and the parameter is inverse proprotion to the size of the result point clouds. The 4th parameter controls the number of smaller pieces. Tipically, the following method works well when the size of the result point cloud files is around 80KB. Make sure that the input files and the executable files are in the same directory. Output: The output of execution is a set of NCD files naming '_{i}_B.ncd' ({i}=1,2,...) as well as two list files 'filelist.txt' and 'filelist_tran.txt', where the common suffix of output files is 'B.ncd'.

Global matching step (!3_docking.bat)

In this step, the coarse matching results are generated. Input: The input files consist of the NCD file of the receptor, and the output files of the Cut NCD step. Configure: As the main executable file has user interface, no configuration is needed. Operate: After opening the RegDock.exe file, several operations are needed. Firstly, right click on 'Reference' in the left area, and click on 'Load Receptor Point Cloud' in the popup menu. Then, select the NCD file of receptor, i.e. 'A.ncd'. Secondly, right click on 'A.ncd' in the left area, and click on 'Enhanced volume invariant'. Thirdly, find 'Global Matching'->'GlobalFile' in the main menu and click it. Choose the file 'filelist.txt' generated in Cut NCD step. Output: The output of execution is a set of rigid body motion files naming '_{i}_B.ncd.tran' ({i}=1,2,...), where the input ligand files are '_{i}_B.ncd'.

Clustering step (!4_cluster.bat)

In this step, the coarse matching results are clustered. Input: The input files are the output files of the Global matching step. Configure: The first parameter is the file 'filelist_tran.txt' generated in Cut NCD step. The second parameter is the result file name after clustering, i.e. 'B_cluster.tran'. Output: The output of execution is a rigid body motion file naming 'B_cluster.tran'.

Local matching step (!5_local.bat)

In this step, the local matching is performed to refine the coarse matching results. Input: The input files are the output files of the Clustering step. Configure: Firstly, we should generate sparser point clouds. The configuration is like the Read PDB step, except that the density is smaller, i.e. the file names of receptor and ligand are 'A.pdb' and 'B.pdb'. The main executable file RegDock.exe is called to do the local matching. Operate: After opening the RegDock.exe, several operations are needed. 1. Right click on 'Reference' in the left area, and click on 'Load Receptor Point Cloud' in the popup menu. Then, select the NCD file of the receptor, i.e. 'A_s.ncd'. Right click on 'A_s.ncd' in the left area, and click on 'Enhanced volume invariant'. 2. Right click on 'Objects' in the left area, and click on 'Load Ligand Point Cloud' in the popup menu. Then, select the NCD file of the ligand, i.e. 'B_s_R.ncd'. Right click on 'B_s_R.ncd' in the left area, and click on 'Enhanced volume invariant'. 3. Find 'Global Matching'->'LoadTrans' in the main menu and click it. Choose the file 'B_cluster.tran' generated in Clustering step. 4. Find 'Local Matching'->'AllLocal' in the main menu and click it. Now the local matching is running. 5. After the local matching finished, find 'Save all: Rigid Body Motion' button in the left area, and click it to save all the refined rigid body motion to a .tran file. Output: The output of execution is the rigid body motion file you saved.

Scoring step (!6_scoring.bat)

In this step, the matching results are scored based on geometric complementary. Input: The input files are the output files of the Clustering step. Configure: The main executable file RegDock.exe is called to do the scoring. Operate: After opening the RegDock.exe, several operations are needed. 1. Right click on 'Reference' in the left area, and click on 'Load Receptor Point Cloud' in the popup menu. Then, select the NCD file of the receptor, i.e. 'A.ncd'. Right click on 'A.ncd' in the left area, and click on 'Enhanced volume invariant'. 2. Right click on 'Objects' in the left area, and click on 'Load Ligand Point Cloud' in the popup menu. Then, select the NCD file of the ligand, i.e. 'B.ncd'. Right click on 'B.ncd' in the left area, and click on 'Enhanced volume invariant'. 3. Find 'Global Matching'->'Scoring' in the main menu and click it. Choose the file 'B_local.tran' generated in the Local matching step. Output: The main output of execution is the scoring file naming 'B.ncd.score'. The first line of 'B.ncd.score' is the number of scored rigid body motions. The rest lines of the result file is in a format like: '(id) (sA) (sB) (vA) (vB) (sA+sB) (csA) (csB) (cvA) (cvB)'. (id) is the 0-based id of the specific rigid body motion. (sA) and (sB) are the surface scores from surface A to B and from B to A respectively. (vA) and (vB) are the volume scores from surface A to B and from B to A respectively. (sA+sB) is the sum of sA and sB. (csA), (csB), (cvA), and (cvB) are the number of vertice/voxels contribute the scores (sA), (sB), (vA), and (vB) respectively.

EXAMPLE

The batch files are all configured to use in the case of the example, assuming that '1PPE_E.pdb' is the receptor and '1PPE_I.pdb' is the ligand. The original PDB file of 1PPE is from Protein Data Bank, and we seperate the two chains to get two PDB files. Extract the two PDB files to the same directory with the executable files to use.

An example consists of 2 PDB files.

Last Modified: 02:00pm, December 27, 2007