Mercurial > repos > galaxyp > openms_idposteriorerrorprobability

annotate test-data/examples/simulation/FASTAProteinAbundanceSampling.py @ 18:6daaa75ccb99 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
author galaxyp
date Sun, 13 Dec 2020 15:03:50 +0000
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
18
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
1 # --------------------------------------------------------------------------
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
2 # OpenMS -- Open-Source Mass Spectrometry
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
3 # --------------------------------------------------------------------------
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
4 # Copyright The OpenMS Team -- Eberhard Karls University Tuebingen,
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
5 # ETH Zurich, and Freie Universitaet Berlin 2002-2020.
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
6 #
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
7 # This software is released under a three-clause BSD license:
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
8 # * Redistributions of source code must retain the above copyright
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
9 # notice, this list of conditions and the following disclaimer.
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
10 # * Redistributions in binary form must reproduce the above copyright
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
11 # notice, this list of conditions and the following disclaimer in the
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
12 # documentation and/or other materials provided with the distribution.
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
13 # * Neither the name of any author or any participating institution
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
14 # may be used to endorse or promote products derived from this software
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
15 # without specific prior written permission.
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
16 # For a full list of authors, refer to the file AUTHORS.
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
17 # --------------------------------------------------------------------------
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
18 # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
19 # AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
20 # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
21 # ARE DISCLAIMED. IN NO EVENT SHALL ANY OF THE AUTHORS OR THE CONTRIBUTING
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
22 # INSTITUTIONS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
23 # EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
24 # PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
25 # OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
26 # WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
27 # OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
28 # ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
29 #
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
30 # --------------------------------------------------------------------------
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
31 # $Maintainer: Chris Bielow $
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
32 # $Authors: Chris Bielow $
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
33 # --------------------------------------------------------------------------
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
34
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
35 import re
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
36 import random
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
37 import math
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
38 import sys
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
39 import argparse
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
40
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
41 ## holds FASTA header + sequence
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
42 class FASTAEntry:
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
43 pass
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
44
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
45 ## grab entries from FASTA file
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
46 def nextEntry(fileobj):
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
47 entry = FASTAEntry()
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
48 entry.header = fileobj.readline()
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
49 entry.sequence = ""
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
50 for line in fileobj:
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
51 if '>' == line[0]:
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
52 yield entry
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
53 entry.header = line
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
54 entry.sequence = ""
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
55 else:
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
56 entry.sequence += line
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
57 yield entry
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
58
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
59
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
60 ## sample abundance from Gaussian in log space
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
61 def sampleAbundance(mu=3, sigma=1):
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
62 return math.exp(random.gauss(mu, sigma))
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
63
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
64
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
65 def main(argv):
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
66 ## we use ArgumentParser, which requires 2.7
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
67 if sys.version_info < (2, 7):
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
68 raise "This script requires python 2.7 or greater"
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
69
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
70 ## add weight filtering functionality if BioPython is available
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
71 try:
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
72 from Bio.SeqUtils.ProtParam import ProteinAnalysis
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
73 has_biopython = 1
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
74 except :
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
75 has_biopython = 0
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
76
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
77
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
78 parser = argparse.ArgumentParser(description='Add abundance to FASTA files.')
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
79 parser.add_argument('infile', type=argparse.FileType('r'), help='Input FASTA file')
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
80 parser.add_argument('outfile', type=argparse.FileType('w'), help='Output FASTA file')
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
81
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
82 parser.add_argument('--mu', dest='mu', action='store', default=3, help='mean of gaussian in log space')
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
83 parser.add_argument('--sigma', dest='sigma', action='store', default=1, help='sd of gaussian in log space')
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
84 parser.add_argument('--sample', dest='sample', action='store', default=0, help='Number of entries to keep (for sampling a bigger FASTA file)')
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
85 parser.add_argument('--random', dest='random', action='store_true', help='Randomly shuffle entries before sampling (only if --sample is given). If not given, the first \'X\' samples are used.')
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
86 if (has_biopython):
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
87 parser.add_argument('--weight_low', dest='weight_low', action='store', default=0, help='minimum molecular weight of protein')
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
88 parser.add_argument('--weight_up', dest='weight_up', action='store', default=0, help='Maximum molecular weight of protein (use 0 for unlimited)')
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
89 else:
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
90 print ("Warning: protein weight filtering not supported, as BioPython module is not installed.")
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
91
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
92 ## argument parsing
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
93 args = parser.parse_args()
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
94 fileobj = args.infile
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
95 fileoutobj = args.outfile
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
96 sample_size = int(args.sample)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
97 sample_random = bool(args.random)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
98 if (has_biopython):
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
99 weight_low = float(args.weight_low)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
100 weight_up = float(args.weight_up)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
101 if (weight_up <= 0): weight_up = sys.float_info.max
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
102
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
103
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
104 ## list of final entries
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
105 fasta_entries = []
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
106
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
107 for entry in nextEntry(fileobj):
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
108 header = entry.header
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
109 ## check if it contains 'intensity'?
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
110 rep = re.compile(r"\[# *(.*) *#\]")
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
111 m = rep.search(header)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
112 header_new = ""
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
113 other = []
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
114 if (m):
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
115 header_new = header.replace(m.group(0), "") ## delete meta
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
116 for element in m.group(1).split(','):
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
117 #print "element:", element
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
118 if (element.find("intensity") == -1):
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
119 other.append(element)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
120 else:
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
121 header_new = header ## nothing to replace
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
122
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
123 ## create new metainfo array
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
124 i = "intensity=" + str(sampleAbundance(float(args.mu), float(args.sigma)))
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
125 other.append(i)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
126
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
127 entry.header = header_new.rstrip() + "[# " + (", ").join(other) + " #]"
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
128
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
129 if (has_biopython):
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
130 sequence = "".join(entry.sequence.split("\n"))
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
131 ##
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
132 ## BioPython does not like some AA letters - they need replacement
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
133 ##
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
134 ## replace "U" (Selenocystein) with "C" (Cystein)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
135 sequence = sequence.replace("U","C")
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
136 ## replace "X" (unknown) with "P" (Proline) [arbitrary choice - but weight of 115 is very close to averagine]
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
137 sequence = sequence.replace("X","P")
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
138 ## replace "B" (Asparagine or aspartic acid) with "N" (Asparagine)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
139 sequence = sequence.replace("B","N")
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
140 ## replace "Z" (Glutamine or glutamic acid) with "Q" (Glutamine)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
141 sequence = sequence.replace("Z","Q")
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
142 ## replace "Z" (Glutamine or glutamic acid) with "Q" (Glutamine)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
143 sequence = sequence.replace("Z","Q")
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
144 ## replace "J" (Leucine or Isoleucine) with "L" (Leucine)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
145 sequence = sequence.replace("J","L")
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
146 analysed_seq = ProteinAnalysis(sequence)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
147 weight = analysed_seq.molecular_weight()
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
148 if (not(weight_low <= weight and weight <= weight_up)):
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
149 continue
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
150
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
151
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
152 fasta_entries.append(entry.header + "\n" + entry.sequence)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
153
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
154 ## only read to sample size (the rest is thrown away anyways)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
155 if (sample_size > 0 and not(sample_random)):
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
156 if (len(fasta_entries) >= sample_size):
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
157 break
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
158
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
159
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
160 ## select subset (if required)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
161 if (sample_size > 0):
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
162 indices = range(0,len(fasta_entries))
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
163 ## random sampling only makes sense if we take a subset
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
164 if (sample_random and sample_size < len(fasta_entries)):
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
165 random.shuffle(indices)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
166 indices = [indices[i] for i in range(0,sample_size)]
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
167 fasta_entries = [fasta_entries[i] for i in indices]
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
168
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
169 ## write to file
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
170 for entry in fasta_entries:
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
171 fileoutobj.write(entry)
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
172
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
173 print ("Generated " + str(len(fasta_entries)) + " protein sequences")
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
174
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
175 if __name__ == "__main__":
6daaa75ccb99 "planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/openms commit 6e7368b7f178fbd1f08c28eea1b538add6943a65-dirty"
galaxyp
parents:
diff changeset
176 main(sys.argv)