From 08ad16adbb8e7b1201d0adffe36c42a4bc19c66b Mon Sep 17 00:00:00 2001
From: Tanushree Tunstall <tanu@tunstall.in>
Date: Mon, 30 Nov 2020 14:11:46 +0000
Subject: [PATCH] added chain_extract.py and pdb_chain_extract.py

---
 scripts/my_pdbtools/chain_extract.py     | 24 ++++++++
 scripts/my_pdbtools/pdb_chain_extract.py | 70 ++++++++++++++++++++++++
 2 files changed, 94 insertions(+)
 create mode 100755 scripts/my_pdbtools/chain_extract.py
 create mode 100755 scripts/my_pdbtools/pdb_chain_extract.py

diff --git a/scripts/my_pdbtools/chain_extract.py b/scripts/my_pdbtools/chain_extract.py
new file mode 100755
index 0000000..084a313
--- /dev/null
+++ b/scripts/my_pdbtools/chain_extract.py
@@ -0,0 +1,24 @@
+#!/usr/bin/python3
+
+#=======================================================================
+# TASK: select specified chains from the pdb & save a cropped PDB with
+# the selected chains. Useful for dimer, etc modelling.
+
+# link for saving each chain as a separate file
+# https://stackoverflow.com/questions/11685716/how-to-extract-chains-from-a-pdb-file
+#=======================================================================
+
+from Bio.PDB import PDBParser, PDBIO, Select
+
+
+# Select() Method to return True for every chain in 'chains'
+class ChainExtract(Select):
+    def __init__(self, chain):
+        self.chain = chain
+    
+    def accept_chain(self, chain):
+        #print(dir(chain))
+        if chain.id in self.chain:
+            return 1
+        else:
+            return 0
diff --git a/scripts/my_pdbtools/pdb_chain_extract.py b/scripts/my_pdbtools/pdb_chain_extract.py
new file mode 100755
index 0000000..3e3cf4b
--- /dev/null
+++ b/scripts/my_pdbtools/pdb_chain_extract.py
@@ -0,0 +1,70 @@
+#!/usr/bin/env python3
+
+# Copyright 2020, Tanushree Tunstall
+# This program is distributed under General Public License v. 3.  See the file
+# COPYING for a copy of the license.
+
+__description__ = \
+"""
+chain_extract.py
+
+extract chain/s from pdb and saves selected chains in a pdb file
+"""
+__author__ = "Tanushree Tunstall"
+__date__ = ""
+#=======================================================================
+import os, shutil, sys
+#from pdbtools.helper import cmdline
+from chain_extract import ChainExtract
+from Bio.PDB import PDBParser, PDBIO, Select
+#from Bio.PDB.PDBParser import PDBParser
+
+#io = PDBIO()
+import argparse
+#=======================================================================
+
+def main():
+    """
+    Function to call if run from command line.
+
+    Example use:
+    pdb_chain_extract.py -f <your_pdb_file> -c <chainid1><chainid2> -p <outpath> -o <outfile>
+    Extracts chain 'A' by default.
+    """
+    arg_parser = argparse.ArgumentParser()
+    arg_parser.add_argument('-i', '--pdb_file', help='provide pdb file', default = 'None')
+    arg_parser.add_argument('-c', '--chain', help='chain/s to extract without spaces.', nargs = '+',  default = 'A', type = list)
+    arg_parser.add_argument('-p', '--out_path', help='specify output path',  default = '.', type = str)
+    arg_parser.add_argument('-o', '--out_file', help='specify output filename. Will be used as a prefix to append chain id and pdb file extension',  default = 'pdbfile', type = str)
+    args = arg_parser.parse_args()
+
+    # Extract chains and write each chain as a separate file
+    pdb_file = args.pdb_file
+    print('input pdb file:', pdb_file)
+    
+    # type = list, makes it a list of lists. Hence extracting the list of chains.
+    chains = args.chain[0]
+    #chains = ['A','B','C']
+    print ('user supplied chain:', chains)
+    
+    # output filename and path
+    outpath = args.out_path
+    outfile = args.out_file
+
+    # get structure
+    p = PDBParser(PERMISSIVE=1)       
+    structure = p.get_structure(pdb_file, pdb_file)
+    print('input pdb filename:', structure.get_id())
+
+    my_chains = chains
+    #my_chains = ['G', 'H'] 
+    c_names = ''.join(my_chains)
+    print('Extracting chains:', my_chains)
+    pdb_chains_file = outpath + '/' + outfile + '_' + c_names + '.pdb'
+    io = PDBIO()
+    io.set_structure(structure)
+    io.save(pdb_chains_file, ChainExtract(my_chains))
+
+if __name__ == "__main__":
+    main()
+