def count(json_tar, json_specs, out=sys.stdout):
"""Print a count matrix of the target genes in a given
GO category."""
specs = json.load(vskip(open(json_specs)))
tar = json.load(vskip(open(json_tar)))
# First thing: remove the genes tagged "NA" if present.
NA = tar.pop('NA', None)
# There must be a "total" key of the target dictionary
# that will be used for the counts.
total = tar.pop('total')
# Turn to sets to remove potential duplicate entries.
for key in specs: specs[key] = set(specs[key])
for key in tar: tar[key] = set(tar[key])
# Get all GO-annotated genes.
annotated = set([gene for ls in specs.values() for gene in ls])
result = [] # List of count lists.
# Proteins in line, GO terms in column. Fill by line.
for prot in sorted(tar):
# Count the intersection of GO and targets.
result.append(
[prot] + \
[str(len(specs[GO].intersection(tar[prot]))) for GO in specs] + \
[str(len(annotated.intersection(tar[prot])))]
)
# The last line contains the total genes with a GO term.
result.append(
['total'] + \
[str(len(specs[GO].intersection(total))) for GO in specs] + \
[str(len(annotated.intersection(total)))]
)
# Print the table header.
out.write('\t'.join(specs) + '\ttotal\n')
# Print the table
for line in result:
out.write('\t'.join(line) + '\n')
def parseGeneAssociations(filename, comment_char='!', columns=(2,5)):
assoVersion = ['-- associations version information --']
pairlist = []
for line in vskip(open(filename)):
if line.startswith(comment_char):
assoVersion.append(line[1:].rstrip())
else:
# Parse by specified columns.
items = line.split('\t')
gene = items[columns[0]-1]
# Skip gene with no canonical ID (flanked by '__')
# and lines with the 'NOT' keyword
if gene[:2] == '__' or re.search('NOT', line):
continue
GOterm = items[columns[1]-1]
pairlist.append((GOterm, gene))
return {
'associations': pairlist,
'assoVersion': assoVersion
}
#!/usr/bin/env python
# -*- coding: utf-8 -*-
try:
import json
except ImportError:
import simplejson as json
import sys
import random
from vtrack import vheader, vskip
def shuffle(jsontar):
"""Modify jsontar in place, using the 'total' field."""
sample = random.Random().sample
for prot in jsontar:
jsontar[prot] = sample(jsontar['total'], len(jsontar[prot]))
if __name__ == '__main__':
jsontar = json.load(vskip(open(sys.argv[1])))
shuffle(jsontar)
sys.stdout.write(vheader(*sys.argv))
json.dump(jsontar, sys.stdout, indent=4)
#! /usr/bin/env python
# -*- coding: utf-8 -*-
import sys
import json
import re
from vtrack import vskip
##########################
# get the mapping table #
# GOID GOTERM #
# get the dict_go from G.F
dict_go = json.load(vskip(open(sys.argv[1])))
for i in dict_go.keys():
goID = re.match('^GO:[0-9]+',i).group(0)
goTERM = re.search('\((.*)\)', i).group(1)
sys.stdout.write("%s\t%s\n"%(goID, goTERM))
def JSONtargets(mappingfile, bindingfile):
"""Create a gene target set in JSON format from a gene mapping
file and a discrete binding profile."""
# Read in gene mapping. Skip comment lines and remove stray
# 'chr' sometimes present in chromosome names.
mapping = [
l.rstrip().replace('chr','').split('\t') \
for l in vskip(open(mappingfile, 'r')) \
if l[0] != '#'
]
# Remove the header if present (recognized by 'start' and
# 'end' in third and fourth columns.
if mapping[0][2:4] == ['start','end']: mapping.pop(0)
# Collect TSS, if gene is on +, TSS is on start, else on end.
TSS = {}
for row in mapping:
thisTSS = {
'+': lambda x: (x[1], int(x[2])), # 2nd and 3rd column.
'-': lambda x: (x[1], int(x[3])) # 2nd and 4th column.
}.get(row[4])(row)
# Arrange geneIDs by TSS in a dictionary.
# Example: TSS['FBgn0031208'] = ('2L', 7529)
TSS[row[0]] = thisTSS
# Read in binding data. Skip comment lines and remove
# 'chr' on chromosome names.
binding = [
l.rstrip().replace('chr','').split('\t') \
for l in vskip(open(bindingfile, 'r')) \
if l[0] != '#'
]
# Get feature names and remove (pop) the header.
# Example: features = ['D005', 'D007', ...]
features = binding.pop(0)[4:]
# "all" and "NA" are mutually exclusive lists of genes.
targets = {'total': [], 'NA': []}
for feature in features:
targets[feature] = []
# Collect mapping information (seqname, start, end) and
# binding info (0/1).
mapinfo = {}
bindinfo = {}
for row in binding:
# Example: mapinfo['r5GATC2L00037'] = ('2L', 5301, 6026)
mapinfo[row[0]] = (row[1], int(row[2]), int(row[3]))
# Example: bindinfo['r5GATC2L00037'] = [0,0,1,...]
bindinfo[row[0]] = row[4:]
# Get the closest feature to TSS.
close_elt = get_closest(TSS, mapinfo, dist = dist)
for geneID in close_elt:
if dist(TSS[geneID], mapinfo[close_elt[geneID]]) > MAXDIST:
# The gene is too far. Push it to NA.
targets.get('NA').append(geneID)
else:
targets.get('total').append(geneID)
# The gene gets the status of the binding element closest
# to its TSS.
for feature in [
feat for (feat, yes) in \
# Example: [('D005', 0), ('D007', 0), ...]
zip(features, bindinfo[close_elt[geneID]]) \
if yes == '1'
]:
targets.get(feature).append(geneID)
# Print the version tracking header and the JSON data.
sys.stdout.write(vheader(*sys.argv))
json.dump(targets, sys.stdout, indent=4)
If the FBgn is not in the table, it is flanked by '__'
which is something to grep for.
"""
from __future__ import with_statement
import re
import sys
from vtrack import vheader, vskip
canonID = {}
# Read-in the lookup table in a dict.
with open(sys.argv[1]) as lookup:
for line in vskip(lookup):
canonID.update([line.rstrip().split('\t')])
def to_canonID(FBmatch):
return canonID.get(
FBmatch.group(),
'__' + FBmatch.group()
)
# Write the vheader.
sys.stdout.write(vheader(*sys.argv))
# Read-in arg file and update FBgn line by line.
with open(sys.argv[2]) as argfile:
for line in argfile:
sys.stdout.write(
re.sub('FBgn[0-9]{7}', to_canonID, line)