diff scoreProfiles.py @ 1:16c9fccf550d draft

planemo upload
author estrain
date Sat, 02 Dec 2017 13:42:43 -0500
parents
children a98ed567ea48
line wrap: on
line diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/scoreProfiles.py	Sat Dec 02 13:42:43 2017 -0500
@@ -0,0 +1,62 @@
+#!/usr/bin/env
+
+import sys
+import glob 
+from decimal import Decimal
+
+allHash = {}
+
+# Read in Profile Database
+profiles = open(sys.argv[1],"r")
+
+# Minimum mean percent coverage for reporting profile
+min_per=float(sys.argv[2])
+# Maximum mean mismatch for reporting profile
+max_mis=float(sys.argv[3])
+
+# Read in Allele Scores
+# Scores should be in srts2*.scores file
+# Column 0:Allele, 1:Score, 2:Avg Depth, 5:% Coverage, 7:Mismatches, 8:Indels
+scoreFile = open(glob.glob("srst2*.scores")[0],"r")
+scoreFile.readline()
+
+for line in scoreFile.readlines() :
+  els = line.split("\t")
+  vals = els[0].split("_")
+  allHash.update({els[0]:line})
+  
+
+# Allele names in first row
+als = profiles.readline().rstrip()
+filehead = als + str("\tMean_Score\tMean_Depth\tMean_%_Coverage\tTotal_Mismatches\tTotal_Indels")
+print(filehead)
+
+genes = als.split("\t")
+
+for line in profiles.readlines() :
+  line = line.rstrip()
+  els = line.split("\t")
+  alleles = [] 
+  for i in range(1,len(genes)) :
+    alleles.append(genes[i] + "_" + els[i])  
+  mscore=0
+  mdepth=0
+  mcover=0
+  mmisma=0
+  mindel=0
+  for i in alleles :
+    if i in allHash :
+      vals=str(allHash[i]).split("\t")
+      mscore+=float(vals[1])
+      mdepth+=float(vals[2])
+      mcover+=float(vals[5])
+      mmisma+=int(vals[7])
+      mindel+=int(vals[8])
+
+  mscore=round(Decimal(mscore/(len(genes)-1)),5)
+  mdepth=round(Decimal(mdepth/(len(genes)-1)),2)
+  mcover=round(Decimal(mcover/(len(genes)-1)),2)
+
+  if mmisma<=max_mis and mcover>=min_per :
+    print(line + "\t" + str(mscore) + "\t" + str(mdepth) + "\t" + str(mcover) + "\t" + str(mmisma) + "\t" + str(mindel)) 
+