def blastp2geneMap(blastFiles,organism1,taxID1,organism2,taxID2,gene2refseq,topNum,IDtype='protein'):
"""
This function transfers blastp results into geneID mapping results
* blastFiles: a list of 2 way blastp tabular result files. eg: [blast1.txt,blast2.txt]
* organism1: string. the 1st organism. eg: 'cho'
* organism2: string. the 2nd organism. eg: 'human'
* gene2refseq: filename. eg: 'gene2refseq'
returns a list of four files. For first 2: each with two columns of gene ID mapping.
For thre rest 2: gene2refseq files for each organism
"""
# extract gene, accession, protein mapping
if IDtype == 'protein':
columnNum = [2,6,7,16]
else:
columnNum = [2,4,5,16]
org1ref = extract_from_gene2ref(gene2refseq,taxID1,organism1,columnNum)
org2ref = extract_from_gene2ref(gene2refseq,taxID2,organism2,columnNum)
result = []
switches = ['False','True']
for blast,switch in zip(blastFiles,switches):
# extract protein id map
pr_id_map = extract_blast_ID_map(blast,topNum,switch) # pr_id_map: cho2human.top1.txt
# protein id mapping to gene id mapping
gene_id_map = mRNA_prIDMap2geneIDMap(pr_id_map,org1ref,org2ref,IDtype) # cho2human.top1.gene.txt
# get unique line of mapping
uniqFile = gene_id_map[:-3] + 'uniqline.txt' # cho2human.top1.gene.uniqline.txt
interFile = gene_id_map[:-3] + 'inter.txt'
# # # get unique id mappings (each line is unique, but genes are not unique)
if switch == 'True':
cmd1 = ('awk -F $\'\\t\' \'BEGIN {OFS} {printrow}\' {input} > {output}').format(
OFS='{FS=\"\\t\"; OFS=FS}',printrow='{print $1,$2}',input=gene_id_map,output=interFile)
cmd2 = ('sort -k2,2 -k1,1 {input} | uniq > {output}').format(input=interFile,output=uniqFile)
else:
cmd1 = ('awk -F $\'\\t\' \'BEGIN {OFS} {printrow}\' {input} > {output}').format(
OFS='{FS=\"\\t\"; OFS=FS}',printrow='{print $1,$2}',input=gene_id_map,output=interFile)
cmd2 = ('sort -k1,1 -k2,2 {input} | uniq > {output}').format(input=interFile,output=uniqFile)
subprocess.call(cmd1,shell=True)
subprocess.call(cmd2,shell=True)
subprocess.call(('rm {inter}').format(inter=interFile),shell=True)
# unique gene ID in 1st column
uniq = uniq1stGene(uniqFile) # cho2human.top1.gene.uniqline.uniq1stgene.txt
result.append(uniq)
result.extend([org1ref,org2ref])
return result
def blastp2geneMap(blastFiles,
organism1,
taxID1,
organism2,
taxID2,
gene2refseq,
topNum,
IDtype='protein'):
"""
This function transfers blastp results into geneID mapping results
* blastFiles: a list of 2 way blastp tabular result files. eg: [blast1.txt,blast2.txt]
* organism1: string. the 1st organism. eg: 'cho'
* organism2: string. the 2nd organism. eg: 'human'
* gene2refseq: filename. eg: 'gene2refseq'
returns a list of four files. For first 2: each with two columns of gene ID mapping.
For thre rest 2: gene2refseq files for each organism
"""
# extract gene, accession, protein mapping
if IDtype == 'protein':
columnNum = [2, 6, 7, 16]
else:
columnNum = [2, 4, 5, 16]
org1ref = extract_from_gene2ref(gene2refseq, taxID1, organism1, columnNum)
org2ref = extract_from_gene2ref(gene2refseq, taxID2, organism2, columnNum)
result = []
switches = ['False', 'True']
for blast, switch in zip(blastFiles, switches):
# extract protein id map
pr_id_map = extract_blast_ID_map(
blast, topNum, switch) # pr_id_map: cho2human.top1.txt
# protein id mapping to gene id mapping
gene_id_map = mRNA_prIDMap2geneIDMap(pr_id_map, org1ref, org2ref,
IDtype) # cho2human.top1.gene.txt
# get unique line of mapping
uniqFile = gene_id_map[:-3] + 'uniqline.txt' # cho2human.top1.gene.uniqline.txt
interFile = gene_id_map[:-3] + 'inter.txt'
# # # get unique id mappings (each line is unique, but genes are not unique)
if switch == 'True':
cmd1 = (
'awk -F $\'\\t\' \'BEGIN {OFS} {printrow}\' {input} > {output}'
).format(OFS='{FS=\"\\t\"; OFS=FS}',
printrow='{print $1,$2}',
input=gene_id_map,
output=interFile)
cmd2 = ('sort -k2,2 -k1,1 {input} | uniq > {output}').format(
input=interFile, output=uniqFile)
else:
cmd1 = (
'awk -F $\'\\t\' \'BEGIN {OFS} {printrow}\' {input} > {output}'
).format(OFS='{FS=\"\\t\"; OFS=FS}',
printrow='{print $1,$2}',
input=gene_id_map,
output=interFile)
cmd2 = ('sort -k1,1 -k2,2 {input} | uniq > {output}').format(
input=interFile, output=uniqFile)
subprocess.call(cmd1, shell=True)
subprocess.call(cmd2, shell=True)
subprocess.call(('rm {inter}').format(inter=interFile), shell=True)
# unique gene ID in 1st column
uniq = uniq1stGene(
uniqFile) # cho2human.top1.gene.uniqline.uniq1stgene.txt
result.append(uniq)
result.extend([org1ref, org2ref])
return result
def DB4unOverlap(unmapGeneIDs, org1ref, org2ref, blastFiles, topPrNum,
topGeneNum):
"""
This function tries to find why the 2wayblastP result don't have overlapped mapping ids.
It builds a database file that has all gene ids mappings from both sides and then we can check whethe there
are some overlapps
* unmapGeneIDs: filename. gene ids that don't have overlapped mapping results.
* or
"""
# -------------- 1. get the protein ids of unoverlapped gene ids ---------------
# org1ref = '/data/shangzhong/CHO2Mouse/2wayBlastPresult/141026gene2refseq.cho.txt'
# org2ref = '/data/shangzhong/CHO2Mouse/2wayBlastPresult/141026gene2refseq.mouse.txt'
# # -------------- 2. get top 5 protein mappings ---------------------------------
# blastFiles = ['/data/shangzhong/CHO2Mouse/2wayBlastPresult/all/cho2mouse.txt','/data/shangzhong/CHO2Mouse/2wayBlastPresult/all/mouse2cho.txt']
indexFile = []
switches = ['False', 'True']
for blast, switch in zip(blastFiles, switches):
pr_id_map = extract_blast_ID_map(
blast, topPrNum, switch) # pr_id_map: cho2human.top5.txt
# protein id mapping to gene id mapping
gene_id_map = mRNA_prIDMap2geneIDMap(
pr_id_map, org1ref, org2ref, switch,
IDtype='protein') # cho2human.top5.gene.txt
# delete the consecutive repeated lines, for lines repeated but are seperated by other lines, they will retain
uniq_id_map = gene_id_map[:-3] + 'uniq.txt' # cho2human.top5.gene.uniq.txt
cmd = ('rev {input} | uniq -f 2 | rev > {output}').format(
input=gene_id_map, output=uniq_id_map)
subprocess.call(cmd, shell=True)
# index the id mapping, this is for one id in org1 has multiple ids in org2 mapped to
index_map = indexUniqline(uniq_id_map,
switch) # cho2human.top5.gene.uniq.index.txt
# sort index file
sort_map = index_map[:-3] + 'sort.txt' # cho2human.top5.gene.uniq.index.sort.txt
if switch == 'False': # sort based on 1st column, then on 2nd column
cmd = ('sort -k1,1n -k2,2n {input} > {output}').format(
input=index_map, output=sort_map)
else:
cmd = ('sort -k2,2n -k1,1n {input} > {output}').format(
input=index_map, output=sort_map)
subprocess.call(cmd, shell=True)
#sort_map = '/data/shangzhong/CHO2Human/2wayBlastPresult/top5/cho2human.top250.gene.uniq.index.sort.txt'
# get uniq first two ids.
uniqline_map = uniqFirst2Col(
sort_map) # cho2human.top5.gene.uniq.index.sort.uniqline.txt
# sort by 1rs and 3rd columns for cho2human, and 2nd and 3rd columns for human2cho
sortbyindex_map = uniqline_map[:-3] + 'index.txt' # cho2human.top5.gene.uniq.index.sort.uniqline.index.txt
if switch == 'False':
cmd = ('sort -k1,1n -k5,5n {input} > {output}').format(
input=uniqline_map, output=sortbyindex_map)
else:
cmd = ('sort -k2,2n -k5,5n {input} > {output}').format(
input=uniqline_map, output=sortbyindex_map)
subprocess.call(cmd, shell=True)
final_index = indexUniqline(
sortbyindex_map, switch
) # cho2human.top5.gene.uniq.index.sort.uniqline.index.index.txt
indexFile.append(final_index)
return indexFile
# -------------- 3. get unoverlapped proteins into a list ---------------
# get unonverlapped ids
unmapGeneIDs = '/data/shangzhong/CHO2Mouse/finalresult/CHO2Mouse_nonOverlap.txt'
geneIds = {}
with open(unmapGeneIDs, 'r') as inputfile:
for line in inputfile:
geneIds[line[:-1]] = [[] for i in range(4)]
# -------------- 4. merge two index files into one ---------------------------------
# indexFile = ['/data/shangzhong/CHO2Mouse/2wayBlastPresult/all/cho2mouse.top250.gene.uniq.index.sort.uniqline.index.index.txt',
# '/data/shangzhong/CHO2Mouse/2wayBlastPresult/all/mouse2cho.top250.gene.uniq.index.sort.uniqline.index.index.txt']
topGeneNum = 5
res = open(indexFile[0], 'r')
for line in res:
item = line[:-1].split('\t')
if item[0] in geneIds:
if int(item[4]) > topGeneNum:
continue
else:
geneIds[item[0]][1].append(item[1])
geneIds[item[0]][0].append(item[4])
else:
continue
res = open(indexFile[1], 'r')
for line in res:
item = line[:-1].split('\t')
if item[0] in geneIds:
if int(item[4]) > topGeneNum:
continue
else:
geneIds[item[0]][2].append(item[1])
geneIds[item[0]][3].append(item[4])
else:
continue
outputfile = indexFile[0][:-22] + 'nonmap.txt' # cho2human.top5.nonmap.txt
output = open(outputfile, 'w')
for key in geneIds:
outline = (
'{key}\t{cho2human}\t{human2cho}\n-\t{cho2humanIndex}\t{human2choIndex}\n'
).format(key=key,
cho2human=','.join(geneIds[key][1]),
human2cho=','.join(geneIds[key][2]),
human2choIndex=','.join(geneIds[key][3]),
cho2humanIndex=','.join(geneIds[key][0]))
output.write(outline)
output.close()
print 'done'
def DB4unOverlap(unmapGeneIDs,org1ref,org2ref,blastFiles,topPrNum,topGeneNum):
"""
This function tries to find why the 2wayblastP result don't have overlapped mapping ids.
It builds a database file that has all gene ids mappings from both sides and then we can check whethe there
are some overlapps
* unmapGeneIDs: filename. gene ids that don't have overlapped mapping results.
* or
"""
# -------------- 1. get the protein ids of unoverlapped gene ids ---------------
# org1ref = '/data/shangzhong/CHO2Mouse/2wayBlastPresult/141026gene2refseq.cho.txt'
# org2ref = '/data/shangzhong/CHO2Mouse/2wayBlastPresult/141026gene2refseq.mouse.txt'
# # -------------- 2. get top 5 protein mappings ---------------------------------
# blastFiles = ['/data/shangzhong/CHO2Mouse/2wayBlastPresult/all/cho2mouse.txt','/data/shangzhong/CHO2Mouse/2wayBlastPresult/all/mouse2cho.txt']
indexFile = []
switches = ['False','True']
for blast,switch in zip(blastFiles,switches):
pr_id_map = extract_blast_ID_map(blast,topPrNum,switch) # pr_id_map: cho2human.top5.txt
# protein id mapping to gene id mapping
gene_id_map = mRNA_prIDMap2geneIDMap(pr_id_map,org1ref,org2ref,switch,IDtype='protein') # cho2human.top5.gene.txt
# delete the consecutive repeated lines, for lines repeated but are seperated by other lines, they will retain
uniq_id_map = gene_id_map[:-3] + 'uniq.txt' # cho2human.top5.gene.uniq.txt
cmd = ('rev {input} | uniq -f 2 | rev > {output}').format(input=gene_id_map,output=uniq_id_map)
subprocess.call(cmd,shell=True)
# index the id mapping, this is for one id in org1 has multiple ids in org2 mapped to
index_map = indexUniqline(uniq_id_map,switch) # cho2human.top5.gene.uniq.index.txt
# sort index file
sort_map = index_map[:-3] + 'sort.txt' # cho2human.top5.gene.uniq.index.sort.txt
if switch == 'False': # sort based on 1st column, then on 2nd column
cmd = ('sort -k1,1n -k2,2n {input} > {output}').format(input=index_map,output=sort_map)
else:
cmd = ('sort -k2,2n -k1,1n {input} > {output}').format(input=index_map,output=sort_map)
subprocess.call(cmd,shell=True)
#sort_map = '/data/shangzhong/CHO2Human/2wayBlastPresult/top5/cho2human.top250.gene.uniq.index.sort.txt'
# get uniq first two ids.
uniqline_map = uniqFirst2Col(sort_map) # cho2human.top5.gene.uniq.index.sort.uniqline.txt
# sort by 1rs and 3rd columns for cho2human, and 2nd and 3rd columns for human2cho
sortbyindex_map = uniqline_map[:-3] + 'index.txt' # cho2human.top5.gene.uniq.index.sort.uniqline.index.txt
if switch == 'False':
cmd = ('sort -k1,1n -k5,5n {input} > {output}').format(input=uniqline_map,output=sortbyindex_map)
else:
cmd = ('sort -k2,2n -k5,5n {input} > {output}').format(input=uniqline_map,output=sortbyindex_map)
subprocess.call(cmd,shell=True)
final_index = indexUniqline(sortbyindex_map,switch) # cho2human.top5.gene.uniq.index.sort.uniqline.index.index.txt
indexFile.append(final_index)
return indexFile
# -------------- 3. get unoverlapped proteins into a list ---------------
# get unonverlapped ids
unmapGeneIDs = '/data/shangzhong/CHO2Mouse/finalresult/CHO2Mouse_nonOverlap.txt'
geneIds = {}
with open(unmapGeneIDs,'r') as inputfile:
for line in inputfile:
geneIds[line[:-1]] = [[] for i in range(4)]
# -------------- 4. merge two index files into one ---------------------------------
# indexFile = ['/data/shangzhong/CHO2Mouse/2wayBlastPresult/all/cho2mouse.top250.gene.uniq.index.sort.uniqline.index.index.txt',
# '/data/shangzhong/CHO2Mouse/2wayBlastPresult/all/mouse2cho.top250.gene.uniq.index.sort.uniqline.index.index.txt']
topGeneNum = 5
res = open(indexFile[0],'r')
for line in res:
item = line[:-1].split('\t')
if item[0] in geneIds:
if int(item[4]) > topGeneNum:
continue
else:
geneIds[item[0]][1].append(item[1])
geneIds[item[0]][0].append(item[4])
else:
continue
res = open(indexFile[1],'r')
for line in res:
item = line[:-1].split('\t')
if item[0] in geneIds:
if int(item[4]) > topGeneNum:
continue
else:
geneIds[item[0]][2].append(item[1])
geneIds[item[0]][3].append(item[4])
else:
continue
outputfile = indexFile[0][:-22] + 'nonmap.txt' # cho2human.top5.nonmap.txt
output = open(outputfile,'w')
for key in geneIds:
outline = ('{key}\t{cho2human}\t{human2cho}\n-\t{cho2humanIndex}\t{human2choIndex}\n').format(key=key,
cho2human=','.join(geneIds[key][1]),human2cho=','.join(geneIds[key][2]),
human2choIndex=','.join(geneIds[key][3]),cho2humanIndex=','.join(geneIds[key][0]))
output.write(outline)
output.close()
print 'done'