对于有残基缺失的晶体结构往往采用同源建模的方式补全，但常规的同源建模方法往往造成非缺失区域残基的挪动，有时我们并不想出现这种状况，尤其是涉及到多个截短体拼合的情况，这时就需要用到约束性同源建模的方法，只对缺失区域补全而尽可能少地改动非缺失区域。效果图如下所示：

1、准备建模序列文件

>P1;Target
sequence:Target:::::::0.00: 0.00
TCFSGDDDSGLNLGFSTSFAEYNEFDGGEKALIGFSTSFAEFDAEAGDSDEQPIFPQHKLTCFS*

前两行格式严格遵守！！！

保存文件名为Target.ali

将模板结构文件存储为temp.pdb文件（单链，除去水和离子）

以下是从pdb文件中提取fasta序列的脚本：

from Bio import PDB
from Bio.PDB import PDBParser, PDBIO
from Bio.PDB.Polypeptide import PPBuilder
from Bio.Seq import Seq
from Bio.SeqRecord import SeqRecord
import sys

def extract_fasta_from_pdb(pdb_file, type):
    parser = PDBParser(QUIET=True)
    if type == "pdb":
        structure = parser.get_structure('PDB', pdb_file)
    if type == "pdb.gz":
        with gzip.open(pdb_file, 'rb') as f_in:
            with open('temp.pdb', 'wb') as f_out:
                f_out.write(f_in.read())
        structure = parser.get_structure('PDB', 'temp.pdb')
    ppb = PPBuilder()
    fasta_records = []
    for model in structure:
        for chain in model:
            polypeptides = ppb.build_peptides(chain)
            for polypeptide in polypeptides:
                sequence = str(polypeptide.get_sequence())
                record = SeqRecord(Seq(sequence), id=chain.id)
                fasta_records.append(record)
    try:
        os.remove('temp.pdb')
    except:
        pass
    return fasta_records
    
def main():
    file = open("out.fasta", "w")
    sys.stdout = file
    print(">temp")
    pdb_file = "./temp.pdb"
    s = extract_fasta_from_pdb(pdb_file, "pdb")
    for record in s:
        print(record.seq)
    
    file.close()
    
if __name__ == "__main__":
    main()

2、与模板蛋白结构进行序列比对

准备align2d.py脚本文件：

from modeller import *

env = Environ()
aln = Alignment(env)
# temp是模板结构名称，C指的是C链
mdl = Model(env, file='temp', model_segment=('FIRST:C','LAST:C'))	
aln.append_model(mdl, align_codes='temp', atom_files='temp.pdb')	
aln.append(file='Target.ali', align_codes='Target')
aln.align2d(max_gap_length=50)
aln.write(file='Target-temp.ali', alignment_format='PIR')	
aln.write(file='Target-temp.pap', alignment_format='PAP')	

执行后，当前路径下生成两个文件Target-temp.ali,Target-temp.pap。

3、 建模

准备loop_modeling.py脚本文件：

from modeller import *
from modeller.automodel import *    

log.verbose()
env = Environ()

# directories for input atom files
env.io.atom_files_directory = ['.', 'temp.pdb']

class MyModel(AutoModel):
    def select_atoms(self):
        #新建模蛋白链编码为A，编号从1开始，因此缺失残基编号要相应挪动
        #编号最好包含首尾各一个残基
        return Selection(self.residue_range('5:A', '13:A'))

a = MyModel(env, alnfile = 'Target-temp.ali',
            knowns = 'temp', sequence = 'Target')
a.starting_model= 1
a.ending_model  = 1

a.make()

4、如何方便地确定新建模型中插入loop的编号

准备脚本文件renumber_pro.py:

import sys
from optparse import OptionParser

def initialize_options() :
    parser = OptionParser();
    parser.add_option( "-p", "--pdb", dest="pdbname", type="string", help="The input pdb that should be renumbered" );
    parser.add_option( "-o", "--output", dest="output", type="string", help="The output file to write the renumbered pdb to; stdout if not given", default="" );
    parser.add_option( "-a", "--chainA_begin", dest="chAbegin", type="int", help="The first residue in chain A", default=1 );
    parser.add_option( "-b", "--chainB_begin", dest="chBbegin", type="int", help="The first residue in chain B", default=1 );
    parser.add_option( "-c", "--chainC_begin", dest="chCbegin", type="int", help="The first residue in chain C", default=1 );
    parser.add_option( "-d", "--chainD_begin", dest="chDbegin", type="int", help="The first residue in chain D", default=1 );
    parser.add_option( "-e", "--chainE_begin", dest="chEbegin", type="int", help="The first residue in chain E", default=1 );
    return parser

def remap_resnum_for_line( line, mapping ):
    if line[0:4] != "ATOM": return line
    chain = line[21];
    if chain not in mapping: return line
    resstring = line[22:27]
    resnum, lastresstring = mapping[ chain ]
    if lastresstring == "" or resstring != lastresstring :
        if lastresstring != "" : resnum += 1
        mapping[ chain ] = (resnum, resstring )
    newresstring = str(resnum) + " "
    if len(newresstring) == 2: newresstring = "   " + newresstring
    elif len(newresstring) == 3: newresstring = "  " + newresstring
    elif len(newresstring) == 4: newresstring = " " + newresstring
    return line[0:22] + newresstring + line[27:]

if __name__ == "__main__":

    parser = initialize_options()
    (options, args) = parser.parse_args()

    pdblines = open( options.pdbname ).readlines()
    mapping = {}
    mapping["A"] = (options.chAbegin, "");
    mapping["B"] = (options.chBbegin, "");
    mapping["C"] = (options.chCbegin, "");
    mapping["D"] = (options.chDbegin, "");
    mapping["E"] = (options.chEbegin, "");
    newlines = []
    for line in pdblines:
        newlines.append( remap_resnum_for_line( line, mapping ))
    if options.output == "" :
        for line in newlines :
            print(line, end=' ')
    else :
        open( options.output, "w").writelines(newlines)

执行：

python renumber_pro.py -p temp.pdb -o output.pdb

用pymol打开原模板文件发现原序列在357-365之间的残基缺失

对应到重编号的文件中是5-13
因此对应上述脚本文件loop_modeling.py中对下方内容进行调整：

return Selection(self.residue_range('5:A', '13:A'))