genesort.py

import sys
from glob import glob

if (len(sys.argv)<=1): match = '*n0*'
else: match = sys.argv[1]
file = glob(match)[0]

counts = {}
with open(file) as f:
  print f.readline()
  for line in f:
    gene = line.split("\t")[7]
    case = line.split("\t")[0]
    if gene in counts:
      if case not in counts[gene]: counts[gene].append(case)
    else: counts[gene] = [case]

for gene in sorted(counts, key = lambda gene: len(counts[gene]), reverse=True):
#  if len(counts[gene]) < 2: break
  group = []
  for line in open(file):
    if line.split("\t")[7] == gene:
      group.append(line)
  lines = ''
  cases = []
  dups = []
  noise = 0
  for line in sorted(group, key = lambda x: (x.split("\t")[1], x.split("\t")[2])):
    case = line.split("\t")[0]
    if case in cases:
      if case not in dups:
        dups.append(case)
        noise += 1  # another one bites the dust
    else: cases.append(case)
    lines += line
  print "%s has %s variants from %s cases (%s cases have exactly 1 variant in this gene)" \
        % (gene, len(group), len(set(cases)), len(set(cases))-noise)
  print lines