diff baseline/script_xlsx.py @ 0:8a5a2abbb870 draft default tip

Uploaded
author davidvanzessen
date Mon, 29 Aug 2016 05:36:10 -0400
parents
children
line wrap: on
line diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/baseline/script_xlsx.py	Mon Aug 29 05:36:10 2016 -0400
@@ -0,0 +1,58 @@
+import xlrd
+import argparse
+
+parser = argparse.ArgumentParser()
+parser.add_argument("--input", help="Excel input file containing one or more sheets where column G has the gene annotation, H has the sequence id and J has the sequence")
+parser.add_argument("--ref", help="Reference file")
+parser.add_argument("--output", help="Output file")
+
+args = parser.parse_args()
+
+gene_column = 6
+id_column = 7
+seq_column = 8
+LETTERS = [x for x in "ABCDEFGHIJKLMNOPQRSTUVWXYZ"]
+
+
+refdic = dict()
+with open(args.ref, 'r') as ref:
+	currentSeq = ""
+	currentId = ""
+	for line in ref.readlines():
+		if line[0] is ">":
+			if currentSeq is not "" and currentId is not "":
+				refdic[currentId[1:]] = currentSeq
+			currentId = line.rstrip()
+			currentSeq = ""
+		else:
+			currentSeq += line.rstrip()
+	refdic[currentId[1:]] = currentSeq
+	
+currentSeq = ""
+currentId = ""
+with xlrd.open_workbook(args.input, 'r') as wb:
+	with open(args.output, 'a') as o:
+		for sheet in wb.sheets():
+			if sheet.cell(1,gene_column).value.find("IGHV") < 0:
+				print "Genes not in column " + LETTERS[gene_column] + ", skipping sheet " + sheet.name
+				continue
+			o.write(">>>" + sheet.name + "\n")
+			outputdic = dict()
+			for rowindex in range(1, sheet.nrows):
+				ref = sheet.cell(rowindex, gene_column).value.replace(">", "")
+				if ref in outputdic:
+					outputdic[ref] += [(sheet.cell(rowindex, id_column).value.replace(">", ""), sheet.cell(rowindex, seq_column).value)]
+				else:
+					outputdic[ref] = [(sheet.cell(rowindex, id_column).value.replace(">", ""), sheet.cell(rowindex, seq_column).value)]
+			#print outputdic
+			
+			for k in outputdic.keys():
+				if k in refdic:
+					o.write(">>" + k + "\n")
+					o.write(refdic[k] + "\n")
+					for seq in outputdic[k]:
+						#print seq
+						o.write(">" + seq[0] + "\n")
+						o.write(seq[1] + "\n")
+				else:
+					print k + " not in reference, skipping " + k