|
3
|
1 #!/usr/bin/env python
|
|
|
2
|
|
|
3 # NOTE: This tool provides the functionality of the PIMA filter_varsacn()
|
|
|
4 # function here https://github.com/appliedbinf/pima_md/blob/main/pima.py#L3012
|
|
|
5
|
|
|
6 import argparse
|
|
|
7 import re
|
|
|
8 import subprocess
|
|
|
9 import sys
|
|
|
10
|
|
|
11
|
|
|
12 def run_command(self, command):
|
|
|
13 try:
|
|
|
14 return re.split('\\n', subprocess.check_output(command, shell=True).decode('utf-8'))
|
|
|
15 except Exception:
|
|
|
16 message = 'Command %s failed: exiting...' % command
|
|
|
17 sys.exit(message)
|
|
|
18
|
|
|
19
|
|
|
20 def filter_varscan(varscan_raw, output):
|
|
|
21 cmd = ' '.join(['cat', varscan_raw,
|
|
|
22 '| awk \'(NR > 1 && $9 == 2 && $5 + $6 >= 15)',
|
|
|
23 '{OFS = "\\t";f = $6 / ($5 + $6); gsub(/.*\\//, "", $4);s = $4;gsub(/[+\\-]/, "", s);$7 = sprintf("%.2f%%", f * 100);'
|
|
|
24 'min = 1 / log(length(s) + 2) / log(10) + 2/10;if(f > min){print}}\'',
|
|
|
25 '1>' + output])
|
|
|
26 output = run_command(cmd)
|
|
|
27
|
|
|
28
|
|
|
29 if __name__ == '__main__':
|
|
|
30 parser = argparse.ArgumentParser()
|
|
|
31
|
|
|
32 parser.add_argument('--varscan_raw', action='store', dest='varscan_raw', help='Raw varscan mpileup VCF file')
|
|
|
33 parser.add_argument('--output', action='store', dest='output', help='Output filtered VCF file')
|
|
|
34
|
|
|
35 args = parser.parse_args()
|
|
|
36
|
|
|
37 filter_varscan(args.varscan_raw, args.output)
|
