Homec4science
Diffusion tfbs_annotate_FIMO (master) Browse

R10523/scripts/compute_background_model.py
2e72ce521fd6master

Last Change
Raw File

compute_background_model.py
View Options

#! /usr/bin/env python3
import os
import sys
import subprocess
import shutil
"""
sys.argv[1]: absolute path to window .bed file
"""
input_bedfile = sys.argv[1]
# get username on SCITAS server
process = subprocess.run('echo $USER', shell=True,
stdout=subprocess.PIPE)
username = process.stdout.decode().rstrip()
scratch_path = '/scratch/'+username
# creating /fimo_background dir in root if it does not exist yet, deleting it to reinitialize the analysis if it already exists
fimo_background_path = os.path.join(scratch_path, 'fimo_background')
if os.path.exists(fimo_background_path): # /fimo_background exists -> reset analysis
print('/fimo_background already exists, deleting it')
shutil.rmtree(fimo_background_path)
os.mkdir(fimo_background_path)
# computing the proportion of each base in the windows to be scanned by FIMO.
# -strand option of faCount is required to take both strands into account
base_freqs_output_path = os.path.join(fimo_background_path, 'base_freqs_raw.txt')
base_freqs_status = subprocess.run("../ucsc_bin/twoBitToFa ../data/hg19.2bit -bed={0} stdout | ../ucsc_bin/faCount -summary -strands stdin > {1}".format(input_bedfile, base_freqs_output_path),
shell=True,
stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
print(base_freqs_status.stdout)
# Formatting the background to a usable MEME 0-order Markov background model
import pandas as pd
base_freqs_raw = pd.read_csv(os.path.join(base_freqs_output_path), sep='\t', header=0)
background_path = os.path.join(fimo_background_path, 'background_markov_model.txt')
with open(background_path, 'w') as f:
f.write('#\torder 0\n')
base_freqs_raw[base_freqs_raw['#seq'] == 'prcnt'][['A', 'C', 'G', 'T']].transpose().to_csv(background_path, sep='\t', mode='a', header=False)
c4science · Help