蛋白质结构文件(pdb)中残基的重新编号

Question

嗨

我目前正在制作一个网站，旨在将所有的乳头状瘤病毒信息组合在一个地方。作为我们在公共服务器(例如genbank)上管理所有已知文件的努力的一部分，我遇到的一个问题是许多(~50%)所有已解决的结构没有根据蛋白质进行编号。也就是说，一个亚域是结晶的(氨基酸310-450)，然而结晶学家将其存放为残基1-140。我想知道有没有人知道给整个pdb文件重新编号的方法。我已经找到了重新编号序列(由seqres标识)的方法，但是这不会更新螺旋和板材信息。如果你有任何建议我将不胜感激…

谢谢

Answer 1

我也经常遇到这个问题。在放弃了一个旧的perl脚本之后，我一直在试验一些python。本解决方案假设您已经安装了Biopython、ProDy (http://www.csb.pitt.edu/ProDy/#prody)和EMBOSS (http://emboss.sourceforge.net/)。

我在这里使用了其中一个乳头瘤病毒PDB条目。

from Bio import AlignIO,SeqIO,ExPASy,SwissProt
from Bio.Seq import Seq
from Bio.SeqRecord import SeqRecord
from Bio.Alphabet import IUPAC
from Bio.Emboss.Applications import NeedleCommandline
from prody.proteins.pdbfile import parsePDB, writePDB
import os

oneletter = {
'ASP':'D','GLU':'E','ASN':'N','GLN':'Q',
'ARG':'R','LYS':'K','PRO':'P','GLY':'G',
'CYS':'C','THR':'T','SER':'S','MET':'M',
'TRP':'W','PHE':'F','TYR':'Y','HIS':'H',
'ALA':'A','VAL':'V','LEU':'L','ILE':'I',
}

# Retrieve pdb to extract sequence
# Can probably be done with Bio.PDB but being able to use the vmd-like selection algebra is nice
pdbname="2kpl"
selection="chain A"
structure=parsePDB(pdbname)
pdbseq_str=''.join([oneletter[i] for i in structure.select("protein and name CA and     %s"%selection).getResnames()])
alnPDBseq=SeqRecord(Seq(pdbseq_str,IUPAC.protein),id=pdbname)
SeqIO.write(alnPDBseq,"%s.fasta"%pdbname,"fasta")

# Retrieve reference sequence
accession="Q96QZ7"
handle = ExPASy.get_sprot_raw(accession)
swissseq = SwissProt.read(handle)
refseq=SeqRecord(Seq(swissseq.sequence,IUPAC.protein),id=accession)
SeqIO.write(refseq, "%s.fasta"%accession,"fasta")

# Do global alignment with needle from EMBOSS, stores entire sequences which makes numbering easier
needle_cli = NeedleCommandline(asequence="%s.fasta"%pdbname,bsequence="%s.fasta"%accession,gapopen=10,gapextend=0.5,outfile="needle.out")
needle_cli()
aln = AlignIO.read("needle.out", "emboss")
os.remove("needle.out")
os.remove("%s.fasta"%pdbname)
os.remove("%s.fasta"%accession)

alnPDBseq = aln[0]
alnREFseq = aln[1]
# Initialize per-letter annotation for pdb sequence record
alnPDBseq.letter_annotations["resnum"]=[None]*len(alnPDBseq)
# Initialize annotation for reference sequence, assume first residue is #1
alnREFseq.letter_annotations["resnum"]=range(1,len(alnREFseq)+1)

# Set new residue numbers in alnPDBseq based on alignment
reslist = [[i,alnREFseq.letter_annotations["resnum"][i]] for i in range(len(alnREFseq)) if alnPDBseq[i] != '-']
for [i,r] in reslist:
    alnPDBseq.letter_annotations["resnum"][i]=r

# Set new residue numbers in the structure
newresnums=[i for i in alnPDBseq.letter_annotations["resnum"][:] if i != None]
resindices=structure.select("protein and name CA and %s"%selection).getResindices()
resmatrix = [[newresnums[i],resindices[i]] for i in range(len(newresnums)) ]
for [newresnum,resindex] in resmatrix:  
    structure.select("resindex %d"%resindex).setResnums(newresnum)

writePDB("%s.renumbered.pdb"%pdbname,structure)

Answer 2

我是pdb-tools的维护者-这可能是一个可以帮助你的工具。

我最近修改了应用程序中的residue-renumber脚本，以提供更大的灵活性。它现在可以renumber六面体和特定的链，或者强制残基编号连续，或者只是将用户指定的偏移量添加到所有残基。

如果这对你有帮助，请告诉我。

Answer 3

1. pdb-tools
2. Phenix pdb-tools
3. BioPython或Bio3D

检查第一个-它应该符合您的需求