Mercurial > repos > yating-l > rename_scaffolds

comparison rename.py @ 2:8fdd3e06e1ec draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload commit eb7c6866b218a41d9b22255ee6afb83551a5ff40
author yating-l
date Mon, 25 Jun 2018 15:29:18 -0400
parents 0d13e4410c3d
children 9529a207d704
comparison
equal deleted inserted replaced
1:0d13e4410c3d 2:8fdd3e06e1ec
1 """ 1 """
2 Call rename to rename scaffolds in reference genome so that the sequence names are less than 31 characters. Rename all scaffolds to scaffold_1, scaffold_2, ..., scaffold_N and provide a name mapping file 2 Call rename to rename scaffolds in reference genome so that the sequence names are less than 31 characters. Rename all scaffolds to scaffold_1, scaffold_2, ..., scaffold_N and provide a name mapping file
3 """ 3 """
4 import sys 4 import sys
5 import csv 5 import csv
6 from collections import OrderedDict
7 6
8 def rename(inputfile, outputfile, writer): 7 def rename(inputfile, outputfile, writer):
9 namemap = OrderedDict()
10 with open(outputfile, 'w') as out: 8 with open(outputfile, 'w') as out:
11 with open(inputfile, 'r') as rf: 9 with open(inputfile, 'r') as rf:
12 lines = rf.readlines() 10 lines = rf.readlines()
13 i = 1 11 i = 1
14 for line in lines: 12 for line in lines:
15 if ">" in line: 13 if ">" in line:
16 oldname = line[1:].rstrip() 14 oldname = line[1:].rstrip()
17 newname = "scaffold_" + str(i) 15 newname = "scaffold_" + str(i)
18 line = ">" + newname 16 line = ">" + newname + "\n"
19 i = i+1 17 i = i+1
20 writer.writerow([oldname, newname]) 18 writer.writerow([oldname, newname])
21 #TODO: Add line breaks to chromosome sequences that are in a single line 19 out.write(line)
22 out.write(line.rstrip() + "\n")
23 20
24 def main(): 21 def main():
25 inputfile = str(sys.argv[1]) 22 inputfile = str(sys.argv[1])
26 outputfile = str(sys.argv[2]) 23 outputfile = str(sys.argv[2])
27 indexfile = str(sys.argv[3]) 24 indexfile = str(sys.argv[3])
28 csvfile = open(indexfile, 'w') 25 csvfile = open(indexfile, 'w')
29 fieldnames = ['Original sequence name', 'Renamed sequence name']
30 writer = csv.writer(csvfile) 26 writer = csv.writer(csvfile)
31 writer.writerow(fieldnames)
32 rename(inputfile, outputfile, writer) 27 rename(inputfile, outputfile, writer)
33 28
34 if __name__ == "__main__": 29 if __name__ == "__main__":
35 main() 30 main()
36 31