comparison tools/protein_analysis/tmhmm2.py @ 7:9b45a8743100 draft

Uploaded v0.1.0, which adds a wrapper for Promoter 2.0 (DNA tool) and enables use of Galaxy's <parallelism> tag for SignalP, TMHMM X Promoter wrappers.

6:a290c6d4e658

#!/usr/bin/env python
"""Wrapper for TMHMM v2.0 for use in Galaxy.

This script takes exactly two command line arguments - an input protein FASTA
filename and an output tabular filename. It then calls the standalone TMHMM
v2.0 program (not the webservice) requesting the short output (one line per
protein).

The first major feature is cleaning up the tabular output. The short form raw
output from TMHMM v2.0 looks like this (six columns tab separated):

 gi|2781234|pdb|1JLY|B	len=304 ExpAA=0.01	First60=0.00	PredHel=0	Topology=o

(since TMHMM v2.0 itself is single threaded) by dividing the input FASTA file
into chunks and running multiple copies of TMHMM in parallel. I would normally
use Python's multiprocessing library in this situation but it requires at
least Python 2.6 and at the time of writing Galaxy still supports Python 2.4.

Also tmhmm2 can fail without returning an error code, for example if run on a
64 bit machine with only the 32 bit binaries installed. This script will spot
when there is no output from tmhmm2, and raise an error.
"""
import sys
import os
from seq_analysis_utils import stop_err, split_fasta, run_jobs

FASTA_CHUNK = 500

if len(sys.argv) != 4:
   stop_err("Require three arguments, number of threads (int), input protein FASTA file & output tabular file")
try:
   num_threads = int(sys.argv[1])
except:
   num_threads = 0
if num_threads < 1:
   stop_err("Threads argument %s is not a positive integer" % sys.argv[1])
fasta_file = sys.argv[2]
tabular_file = sys.argv[3]

def clean_tabular(raw_handle, out_handle):
    """Clean up tabular TMHMM output, returns output line count."""
    count = 0
    for line in raw_handle:

        count += 1
    return count

#Note that if the input FASTA file contains no sequences,
#split_fasta returns an empty list (i.e. zero temp files).
fasta_files = split_fasta(fasta_file, tabular_file, FASTA_CHUNK)
temp_files = [f+".out" for f in fasta_files]
jobs = ["tmhmm -short %s > %s" % (fasta, temp)
        for fasta, temp in zip(fasta_files, temp_files)]

def clean_up(file_list):
    for f in file_list:
        if os.path.isfile(f):
            os.remove(f)

if len(jobs) > 1 and num_threads > 1:
    #A small "info" message for Galaxy to show the user.
    print "Using %i threads for %i tasks" % (min(num_threads, len(jobs)), len(jobs))
results = run_jobs(jobs, num_threads)

    if error_level:
        try:
            output = open(temp).readline()
        except IOError:
            output = ""
        clean_up(fasta_files)
        clean_up(temp_files)
        stop_err("One or more tasks failed, e.g. %i from %r gave:\n%s" % (error_level, cmd, output),
                 error_level)
del results
del jobs

for temp in temp_files:
    data_handle = open(temp)
    count = clean_tabular(data_handle, out_handle)
    data_handle.close()
    if not count:
        clean_up(fasta_files)
        clean_up(temp_files)
        stop_err("No output from tmhmm2")
out_handle.close()

clean_up(fasta_files)
clean_up(temp_files)

7:9b45a8743100

#!/usr/bin/env python
"""Wrapper for TMHMM v2.0 for use in Galaxy.

This script takes exactly three command line arguments - number of threads,
an input protein FASTA filename, and an output tabular filename. It then
calls the standalone TMHMM v2.0 program (not the webservice) requesting
the short output (one line per protein).

The first major feature is cleaning up the tabular output. The short form raw
output from TMHMM v2.0 looks like this (six columns tab separated):

 gi|2781234|pdb|1JLY|B	len=304 ExpAA=0.01	First60=0.00	PredHel=0	Topology=o

(since TMHMM v2.0 itself is single threaded) by dividing the input FASTA file
into chunks and running multiple copies of TMHMM in parallel. I would normally
use Python's multiprocessing library in this situation but it requires at
least Python 2.6 and at the time of writing Galaxy still supports Python 2.4.

Note that this is somewhat redundant with job-splitting available in Galaxy
itself (see the SignalP XML file for settings).

Also tmhmm2 can fail without returning an error code, for example if run on a
64 bit machine with only the 32 bit binaries installed. This script will spot
when there is no output from tmhmm2, and raise an error.
"""
import sys
import os
import tempfile
from seq_analysis_utils import stop_err, split_fasta, run_jobs

FASTA_CHUNK = 500

if len(sys.argv) != 4:
   stop_err("Require three arguments, number of threads (int), input protein FASTA file & output tabular file")
try:
   num_threads = int(sys.argv[1])
except:
   num_threads = 1 #Default, e.g. used "$NSLOTS" and environment variable not defined
if num_threads < 1:
   stop_err("Threads argument %s is not a positive integer" % sys.argv[1])
fasta_file = sys.argv[2]
tabular_file = sys.argv[3]

tmp_dir = tempfile.mkdtemp()

def clean_tabular(raw_handle, out_handle):
    """Clean up tabular TMHMM output, returns output line count."""
    count = 0
    for line in raw_handle:

        count += 1
    return count

#Note that if the input FASTA file contains no sequences,
#split_fasta returns an empty list (i.e. zero temp files).
fasta_files = split_fasta(fasta_file, os.path.join(tmp_dir, "tmhmm"), FASTA_CHUNK)
temp_files = [f+".out" for f in fasta_files]
jobs = ["tmhmm -short %s > %s" % (fasta, temp)
        for fasta, temp in zip(fasta_files, temp_files)]

def clean_up(file_list):
    for f in file_list:
        if os.path.isfile(f):
            os.remove(f)
    try:
        os.rmdir(tmp_dir)
    except:
        pass

if len(jobs) > 1 and num_threads > 1:
    #A small "info" message for Galaxy to show the user.
    print "Using %i threads for %i tasks" % (min(num_threads, len(jobs)), len(jobs))
results = run_jobs(jobs, num_threads)

    if error_level:
        try:
            output = open(temp).readline()
        except IOError:
            output = ""
        clean_up(fasta_files + temp_files)
        stop_err("One or more tasks failed, e.g. %i from %r gave:\n%s" % (error_level, cmd, output),
                 error_level)
del results
del jobs

for temp in temp_files:
    data_handle = open(temp)
    count = clean_tabular(data_handle, out_handle)
    data_handle.close()
    if not count:
        clean_up(fasta_files + temp_files)
        stop_err("No output from tmhmm2")
out_handle.close()

clean_up(fasta_files + temp_files)