listGeneFeatures.py

   1 #
   2 #  listGeneFeatures.py
   3 #  ENRAGE
   4 #
   5
   6 import sys
   7 from cistematic.genomes import Genome
   8 from commoncode import getMergedRegions, getFeaturesByChromDict
   9
  10 print "%s: version 1.1" % sys.argv[0]
  11
  12
  13 def main(argv=None):
  14     if not argv:
  15         argv = sys.argv
  16
  17     if len(argv) < 4:
  18         print "usage: python %s genome [acceptFile] gid outfile\n" % argv[0]
  19         sys.exit(1)
  20
  21     genome = argv[1]
  22
  23     if len(argv) == 4:
  24         gid = argv[2]
  25         outfile = argv[3]
  26     else:
  27         acceptFileName = argv[2]
  28         gid = argv[3]
  29         outfile = argv[4]
  30
  31     listGeneFeatures(genome, gid, outfile, acceptFileName)
  32
  33
  34 def listGeneFeatures(genome, gid, outFileName, acceptFileName=""):
  35     hg = Genome(genome)
  36     outfile = open(outFileName, "w")
  37     if acceptFileName:
  38         additionalDict = getMergedRegions(acceptFileName, maxDist = 0, keepLabel = True, verbose = True)
  39     else:
  40         additionalDict = {}
  41
  42     featuresDict = getFeaturesByChromDict(hg, additionalDict, restrictList=[gid])
  43     outfile.write('track name="LOC%s"\n' % gid)
  44
  45     senseDict = {"F": "+",
  46                  "R": "-",
  47                  "+": "+",
  48                  "-": "-"
  49     }
  50
  51     for chrom in featuresDict:
  52         for (start, stop, fgid, sense, ftype) in featuresDict[chrom]:
  53             outfile.write("chr%s\t%d\t%d\t%s\t0\t%s\n" % (chrom, start, stop, ftype, senseDict[sense]))
  54
  55     outfile.close()
  56
  57
  58 if __name__ == "__main__":
  59     main(sys.argv)