跳转到内容

去除细胞器序列

作用:将 contig 比对 NCBI nt,经 GI → taxid → 学名汇总每个物种的命中数,据此识别并去除质体/线粒体来源的 contig。 入口:03.1.blast.sh(内部调用 03.2.get_gi.py、03.3.get_name.py),随后运行 03.4.get_ratio.py 输入:contigs.fa;nt 库 ``{SOFTWARE_ZC}/nt_lib/nt;cutted_nucl_gb.accession2taxid 与 names.dmp(须放在工作目录)。 输出:nt.blast.xml、tiqu_gi.txt、scientific_name.txt、final_result.txt(Name/Hit_reads/percentage1/percentage2,按命中数排序)。 运行:bash 03.1.blast.sh,然后 python 03.4.get_ratio.py 工具:blastn -outfmt 5 -evalue 1e-5 -num_threads 88;Python 仅用标准库(re、collections),以 python 调用。

English

Purpose: BLAST the contigs against NCBI nt, resolve GI → taxid → scientific name and count hits per species to identify and remove plastid/mitochondrial (organelle) contigs. Entry point: 03.1.blast.sh (calls 03.2.get_gi.py and 03.3.get_name.py), after which 03.4.get_ratio.py runs separately Inputs: contigs.fa; nt database ``{SOFTWARE_ZC}/nt_lib/nt; cutted_nucl_gb.accession2taxid and names.dmp (must be in the working directory). Outputs: nt.blast.xml, tiqu_gi.txt, scientific_name.txt, final_result.txt (Name/Hit_reads/percentage1/percentage2, sorted by hit count). Run: bash 03.1.blast.sh, then python 03.4.get_ratio.py Tools: blastn -outfmt 5 -evalue 1e-5 -num_threads 88; Python with the standard library only (re, collections), invoked as python.

以下为该步骤目录中的实际文件(同步快照,与该目录内容一致)。

03.1.blast.sh · Shell · 6 行

在 GitHub 上查看此文件

终端窗口
#3.1 排除长度较长(10Mb)的染色体序列,提取其余序列(可能是质体),比对到nt库(需要从ncbi下载到本地),输出fmt5格式
blastn -query contigs.fa -out nt.blast.xml -db {SOFTWARE_ZC}/nt_lib/nt -outfmt 5 -evalue 1e-5 -num_threads 88
python 03.2.get_gi.py
python 03.3.get_name.py
03.2.get_gi.py · Python · 31 行

在 GitHub 上查看此文件

# -*- coding:utf-8 -*-
import re
from collections import defaultdict
xmlfile=open("nt.blast.xml","r")
outfile=open("tiqu_gi.txt","w")
dict1=defaultdict(list)
for lines in xmlfile:
line=lines.strip()
read_id = re.match('<Iteration_query-def>.*</Iteration_query-def>',line)
Hit_id = re.match('<Hit_id>.*</Hit_id>',line)
Hit_def = re.match('<Hit_def>.*</Hit_def>',line)
if read_id !=None:
read_id=read_id.group()
read_id = read_id.split("<")[1].split(">")[1]
key=read_id
elif Hit_id !=None:
Hit_id = Hit_id.group()
Hit_id = Hit_id.split("<")[1].split(">")[1]
dict1[key].append(Hit_id)
elif Hit_def !=None:
Hit_def = Hit_def.group()
Hit_def = Hit_def.split("<")[1].split(">")[1]
dict1[key].append(Hit_def)
for key in dict1:
outfile.write(key + "\t" + "\t".join(dict1[key])+"\n")
03.3.get_name.py · Python · 42 行

在 GitHub 上查看此文件

# -*- coding:utf-8 -*-
tiqu_gi = open("tiqu_gi.txt","r")
gi2taxid = open("cutted_nucl_gb.accession2taxid","r")
taxid2name = open("names.dmp","r")
get_name = open("scientific_name.txt","w")
from collections import defaultdict
taxid_name_dict={}
for lines in taxid2name:
if "scientific name" in lines:
line = lines.strip().split("|")
taxid = line[0].strip()
name = line[1].strip()
taxid_name_dict[taxid]=name
tiqu_dict=defaultdict(list)
for lines in tiqu_gi:
line = lines.strip().split("\t")
gi = line[1].split("|")[1]
tiqu_dict[gi].append("\t".join(line))
gi_taxid_dict={}
for lines in gi2taxid:
line = lines.strip().split("\t")
GI = line[1]
taxid = line[0]
gi_taxid_dict[GI]=taxid
jiaoji=set(tiqu_dict.keys())&set(gi_taxid_dict.keys())
tax_list=taxid_name_dict.keys()
tiqu_gi = open("tiqu_gi.txt","r")
for lines in tiqu_gi:
line = lines.strip().split("\t")
gi = line[1].split("|")[1]
if gi in jiaoji:
taxid=gi_taxid_dict[gi]
if taxid in tax_list:
get_name.write("\t".join(line)+"\t"+taxid_name_dict[taxid]+"\n")
03.4.get_ratio.py · Python · 24 行

在 GitHub 上查看此文件

# -*- coding:utf-8 -*-
from collections import Counter
scientific_name=open("scientific_name.txt","r")
final_result =open("final_result.txt","w")
name_list_all=[]
for lines in scientific_name:
line = lines.strip().split("\t")
name = line[-1]
name_list_all.append(name)
count_result = Counter(name_list_all)
count_list = list(count_result.items())
count_list.sort(key=lambda x:x[1],reverse=True)
final_result.write("Name\tHit_reads\tpercentage1\tpercentage2\n")
for i in count_list:
name = i[0]
number = i[1]
reads_num = 100000
percentage1 = "%.2f%%"%(100*float(number)/float(reads_num))
percentage2 ="%.2f%%"%(100*float(number)/float(len(name_list_all)))
final_result.write(name+"\t"+str(number)+"\t"+str(percentage1)+"\t"+str(percentage2)+"\n")