Mercurial > repos > saskia-hiltemann > cgatools_v17

annotate tools/cgatools17/vcf_to_listVariants.py @ 1:3a2e0f376f26 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Minor change to tv2vcf.xml to allow for workflow automation
author dgdekoning
date Wed, 21 Oct 2015 10:09:15 -0400
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
1
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
1 #!/usr/bin/env python
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
2
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
3 """[License: GNU General Public License v3 (GPLv3)]
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
4
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
5 This is free software: you can redistribute it and/or modify
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
6 it under the terms of the GNU General Public License as published by
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
7 the Free Software Foundation, either version 3 of the License, or
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
8 (at your option) any later version.
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
9
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
10 This is distributed in the hope that it will be useful,
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
11 but WITHOUT ANY WARRANTY; without even the implied warranty of
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
13 GNU General Public License for more details.
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
14
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
15 You should have received a copy of the GNU General Public License
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
16 along with this program. If not, see <http://www.gnu.org/licenses/>.
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
17
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
18 Documentation as defined by:
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
19 <http://epydoc.sourceforge.net/manual-fields.html#fields-synonyms>
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
20 """
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
21
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
22 import sys,os,os.path,argparse,textwrap,datetime
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
23
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
24 def convert_vcf_to_listVariants(filename_in, filename_out):
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
25 if(filename_out == "-"):
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
26 fh_out = sys.stdout
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
27 else:
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
28 fh_out = open(filename_out,"w")
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
29
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
30 i = 1
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
31
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
32 fh_out.write("variantId\tchromosome\tbegin\tend\tvarType\treference\talleleSeq\txRef\n")
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
33
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
34 with open(filename_in, 'r') as fh_in:
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
35 for line in fh_in:
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
36 line_s = line.strip()
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
37 if((len(line_s) >= 6) and line_s[0] != "#"):
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
38 params = line.split("\t")
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
39
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
40 if(len(params[4]) == 1):# single base substitution
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
41 fh_out.write(str(i)) # id
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
42
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
43 fh_out.write("\t"+params[0]) # chr
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
44 fh_out.write("\t"+str(int(params[1])-1)) # begin
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
45 fh_out.write("\t"+params[1]) # end
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
46 fh_out.write("\tsnp")
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
47 fh_out.write("\t"+params[3]) # reference
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
48 fh_out.write("\t"+params[4]) # alleleSeq
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
49 fh_out.write("\t"+params[2]) # dbsnpid / annotation id
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
50
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
51 fh_out.write("\n")
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
52 i += 1
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
53 # else: # indel...
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
54
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
55 if(filename_out != "-"):
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
56 fh_out.close()
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
57
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
58 if __name__ == "__main__":
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
59 parser = argparse.ArgumentParser()
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
60
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
61 parser.add_argument("-i","--input", help="input file (VCF)")
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
62 parser.add_argument("-o","--output",help="output filename; '-' for stdout",default="-")
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
63
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
64 args = parser.parse_args()
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
65
3a2e0f376f26 Minor change to tv2vcf.xml to allow for workflow automation
dgdekoning
parents:
diff changeset
66 convert_vcf_to_listVariants(args.input,args.output)