Mercurial > repos > davidvanzessen > imgt_loader_igg

comparison imgtconvert.py @ 1:4405849aa053 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author davidvanzessen
date Thu, 06 Mar 2014 09:34:10 -0500
parents f8ba6ab5514b
children 477cab9c32d1
comparison
equal deleted inserted replaced
0:f8ba6ab5514b 1:4405849aa053
1 import pandas as pd 1 import pandas as pd
2 pd.options.mode.chained_assignment = None # default='warn'
2 import re 3 import re
3 import argparse 4 import argparse
4 import os 5 import os
5 6
6 def stop_err( msg, ret=1 ): 7 def stop_err( msg, ret=1 ):
106 107
107 outFrame = outFrame.append(tmp) 108 outFrame = outFrame.append(tmp)
108 109
109 outFrame.columns = [u'ID', u'VDJ Frame', u'Top V Gene', u'Top D Gene', u'Top J Gene', u'CDR1 Seq', u'CDR1 Length', u'CDR2 Seq', u'CDR2 Length', u'CDR3 Seq', u'CDR3 Length', u'CDR3 Seq DNA', u'CDR3 Length DNA', u'Strand', u'CDR3 Found How', u'Functionality', 'V-REGION identity %', 'V-REGION identity nt', 'D-REGION reading frame', 'AA JUNCTION', 'Functionality comment', 'Sequence', 'FR1-IMGT', 'FR2-IMGT', 'FR3-IMGT', 'CDR3-IMGT', 'JUNCTION', 'J-REGION', 'FR4-IMGT', 'P3V-nt nb', 'N1-REGION-nt nb', 'P5D-nt nb', 'P3D-nt nb', 'N2-REGION-nt nb', 'P5J-nt nb', '3V-REGION trimmed-nt nb', '5D-REGION trimmed-nt nb', '3D-REGION trimmed-nt nb', '5J-REGION trimmed-nt nb'] 110 outFrame.columns = [u'ID', u'VDJ Frame', u'Top V Gene', u'Top D Gene', u'Top J Gene', u'CDR1 Seq', u'CDR1 Length', u'CDR2 Seq', u'CDR2 Length', u'CDR3 Seq', u'CDR3 Length', u'CDR3 Seq DNA', u'CDR3 Length DNA', u'Strand', u'CDR3 Found How', u'Functionality', 'V-REGION identity %', 'V-REGION identity nt', 'D-REGION reading frame', 'AA JUNCTION', 'Functionality comment', 'Sequence', 'FR1-IMGT', 'FR2-IMGT', 'FR3-IMGT', 'CDR3-IMGT', 'JUNCTION', 'J-REGION', 'FR4-IMGT', 'P3V-nt nb', 'N1-REGION-nt nb', 'P5D-nt nb', 'P3D-nt nb', 'N2-REGION-nt nb', 'P5J-nt nb', '3V-REGION trimmed-nt nb', '5D-REGION trimmed-nt nb', '3D-REGION trimmed-nt nb', '5J-REGION trimmed-nt nb']
110 111
111 vPattern = re.compile(r"IGHV[1-9]-[0-9ab]+-?[1-9]?") 112 #vPattern = re.compile(r"|TRBV[0-9]{1,2}-?[0-9]?") #mouse
112 dPattern = re.compile(r"IGHD[1-9]-[0-9ab]+") 113 #vPattern = re.compile(r"IGHV[1-9]-[0-9ab]+-?[1-9]?") #human
113 jPattern = re.compile(r"IGHJ[1-9]") 114 vPattern = re.compile(r"IGHV[1-9]-[0-9ab]+-?[1-9]?|TRBV[0-9]{1,2}-?[0-9]?") #mouse and human
115
116 #dPattern = re.compile(r"TRBD1|TRBD2") #mouse
117 #dPattern = re.compile(r"IGHD[1-9]-[0-9ab]+") #human
118 dPattern = re.compile(r"IGHD[1-9]-[0-9ab]+|TRBD1|TRBD2") #mouse and human
119
120
121 #jPattern = re.compile(r"TRBJ[12]-[1-7]") #mouse
122 #jPattern = re.compile(r"IGHJ[1-6]") #human
123 jPattern = re.compile(r"IGHJ[1-6]|TRBJ[12]-[1-7]") #mouse and human
114 124
115 def filterGenes(s, pattern): 125 def filterGenes(s, pattern):
116 if type(s) is not str: 126 if type(s) is not str:
117 return "NA" 127 return "NA"
118 res = pattern.search(s) 128 res = pattern.search(s)