Mercurial > repos > jay > pdaug_peptide_sequence_analysis

annotate PDAUG_Peptide_Core_Descriptors/PDAUG_Peptide_Core_Descriptors.py @ 4:b73ba641e514 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
author jay
date Tue, 12 Jan 2021 18:31:46 +0000
parents e59674e3a391
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
1 from modlamp.core import BaseDescriptor
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
2 from modlamp.descriptors import PeptideDescriptor
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
3 import pandas as pd
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
4 import argparse, os
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
5
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
6 parser = argparse.ArgumentParser()
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
7
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
8 parser.add_argument("-I", "--InFile", required=True, default=None, help="Input file")
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
9 parser.add_argument("-O", "--OutFile", required=True, default=None, help="Output file")
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
10 parser.add_argument("-N", "--Ngrams", required=True, default=None, help="ngrams")
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
11
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
12 args = parser.parse_args()
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
13
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
14 file = open(args.InFile)
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
15 lines = file.readlines()
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
16
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
17 Index = []
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
18 Pep = []
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
19
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
20
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
21 for line in lines:
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
22 if '>' in line:
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
23 Index.append(line.strip('\n'))
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
24 else:
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
25 line = line.strip('\n')
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
26 line = line.strip('\r')
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
27 Pep.append(line)
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
28
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
29 df = pd.DataFrame()
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
30
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
31 for i, l in enumerate(Pep):
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
32
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
33 D = PeptideDescriptor(l)
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
34 D.count_ngrams([int(args.Ngrams)])
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
35
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
36 df1 = pd.DataFrame(D.descriptor, index=["sequence"+str(i),])
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
37 df = pd.concat([df, df1], axis=0)
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
38
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
39 df = df.fillna(0)
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
40 df.to_csv(args.OutFile, sep='\t', index=None)
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
41
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
42
e59674e3a391 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 6f53ad797ec1af02b41510063a86bec7d121abf3"
jay
parents:
diff changeset
43