#!/usr/bin/env python3
__author__ = "Darrin Lemmer"
__version__ = "1.0.0"
__email__ = "dlemmer@tgen.org"
'''
Created on Jun 11, 2014
@author: dlemmer
'''
from nasp import __version__ as nasp_version
import logging
import os
import re
try:
import readline
import glob
[docs] def complete(text, state):
"""
Tab autocomplete for prompts.
Args:
text (str): current user input
state (int): index from 0 to n until the function returns a non-string value
Returns:
str: matching file at the current state index
"""
matches = glob.glob(text+'*')
return matches[state] + os.sep if os.path.isdir(matches[state]) else matches[state]
readline.set_completer_delims('')
readline.parse_and_bind("tab: complete")
readline.set_completer(complete)
except ImportError:
# The readline module is unavailable, tab-autocomplete will not work
pass
def _parse_args():
import argparse
parser = argparse.ArgumentParser(prog="nasp",
description="This is the \"Northern Arizona SNP Pipeline\", version %s" % nasp_version)
parser.add_argument("reference_fasta", nargs="?", default="", help="Path to the reference fasta.")
parser.add_argument("output_folder", nargs="?", default="", help="Folder to store the output files.")
parser.add_argument("--config", help="Path to the configuration xml file.")
return parser.parse_args()
def _expand_path(path):
user_match = re.match('^(~)(.*)$', path)
if user_match:
path = os.path.expanduser(path)
return os.path.abspath(path)
def _create_file_tuple(path):
name = os.path.splitext(os.path.basename(path))[0]
return name, path
def _find_files(path, extension):
file_list = []
for file in os.listdir(path):
is_type = re.search('(.*)(\.%s)$' % extension, file, re.IGNORECASE)
if is_type:
sample_name = is_type.group(1)
file_list.append((sample_name, os.path.join(path, file)))
logging.info((sample_name, os.path.join(path, file)))
return file_list
def _find_executable(application): # This method is not OS-independent. Should work on a better way
import subprocess
executable = ""
try:
executable = subprocess.check_output(["which", application])
except subprocess.CalledProcessError:
try:
executable = subprocess.check_output("find ~ -name %s" % application, shell=True)
except subprocess.CalledProcessError as cpe:
executable = cpe.output
match = re.search('^(.*)\n.*$', str(executable, "utf-8"))
if match:
executable = match.group(1)
return executable
def _find_reads(path):
read_list = []
for file in os.listdir(path):
is_read = re.search('(.*)(\.fastq(?:\.gz)?)$', file, re.IGNORECASE)
if is_read:
sample_name = is_read.group(1)
is_paired = re.search('^(.*)(_[R]?)([12])(.*)$', sample_name, re.IGNORECASE)
if is_paired:
if is_paired.group(3) == '1': # If paired, only process read 1, so we don't double count the pair
sample_name = is_paired.group(1)
read1 = file
read2 = "%s%s2%s%s" % (is_paired.group(1), is_paired.group(2), is_paired.group(4), is_read.group(2))
if os.path.exists(os.path.join(path, read2)):
read = (sample_name, os.path.join(path, read1), os.path.join(path, read2))
read_list.append(read)
logging.info(read)
else:
logging.warning("Cannot find %s, the matching read to %s. Skipping..." % (read2, read1))
else:
read = (sample_name, os.path.join(path, file))
read_list.append(read)
logging.info(read)
return read_list
def _get_bams(cwd):
bam_list = []
response = input("\nDo you have pre-aligned SAM/BAM files you wish to include [N]? ")
if re.match('^[Yy]', response):
path = input("Where are these files located [%s]? " % cwd)
path = _expand_path(path) if path else cwd
logging.info("Looking for sams/bams in %s...", path)
bam_list = _find_files(path, "bam")
bam_list.extend(_find_files(path, "sam"))
return bam_list
def _get_vcfs(cwd):
vcf_list = []
response = input("\nDo you have pre-called VCFfiles you wish to include [N]? ")
if re.match('^[Yy]', response):
path = input("Where are these files located [%s]? " % cwd)
path = _expand_path(path) if path else cwd
logging.info("Looking for vcfs in %s...", path)
vcf_list = _find_files(path, "vcf")
return vcf_list
def _get_external_fastas(cwd, exclude):
fasta_list = []
response = input("\nDo you have fasta files for external genomes you wish to include [Y]? ")
if not re.match('^[Nn]', response):
path = input("Where are these files located [%s]? " % cwd)
path = _expand_path(path) if path else cwd
logging.info("Looking for external fastas in %s...", path)
fasta_list = _find_files(path, "(?:fa|fna|fas|fasta)$")
# remove the reference from the fasta_list if it is in there
for fasta_tuple in fasta_list:
fasta = fasta_tuple[1]
if fasta == exclude:
fasta_list.remove(fasta_tuple)
return fasta_list
def _get_reads(cwd):
read_list = []
response = input("\nDo you have read files you wish to include [Y]? ")
if not re.match('^[Nn]', response):
path = input("Where are these files located [%s]? " % cwd)
path = _expand_path(path) if path else cwd
logging.info("Looking for read files in %s...", path)
read_list = _find_reads(path)
return read_list
def _get_application_path(application):
app_path = _find_executable(application)
if not app_path:
app_path = input("\nUnable to find '%s', please enter the full path to '%s': " % (application, application))
while not os.access(app_path, os.X_OK):
app_path = input(
"\n'%s' either does not exist or you don't have permission to run it, please enter the full path to '%s': " % (
app_path, application))
return app_path
# TODO(jtravis): search CLASSPATH
# NOTE: Wildcard characters are allowed in 'jarfile' as both fnmatch and glob will handle them
def _get_java_path(jarfile):
import fnmatch
paths = ['/usr/share/java/']
paths.extend(os.environ['PATH'].split(os.pathsep))
for path in paths:
match_list = glob.glob(os.path.join(path, jarfile))
if match_list:
return match_list[0]
# Didn't find it in path, check user's home directory
for path, dirs, files in os.walk(os.path.expanduser("~")):
for filename in fnmatch.filter(files, jarfile):
return os.path.join(path, filename)
#Didn't find it there, recursively check current directory
for path, dirs, files in os.walk(os.getcwd()):
for filename in fnmatch.filter(files, jarfile):
return os.path.join(path, filename)
#Let's ask the user
jar_path = input("\nUnable to find '%s', please enter the full path to '%s': " % (jarfile, jarfile))
while not os.access(jar_path, os.R_OK) or not os.path.isfile(jar_path):
# If the user entered a directory path instead of a complete path, check if the file exists in the directory
if os.path.isdir(jar_path):
for path in os.listdir(jar_path):
if path.endswith(os.pathsep + jarfile):
return path
jar_path = input(
"\n'%s' either does not exist or you don't have permission to run it, please enter the full path to '%s': " % (
jar_path, jarfile))
return jar_path
def _get_advanced_settings(app_name, app_path, app_args, job_parms):
response = input("Would you like to set advanced %s settings [N]? " % app_name)
if re.match('^[Yy]', response):
alt_version = input(" Would you like to use an alternate %s version [N]? " % app_name)
if re.match('^[Yy]', alt_version):
path = input(" What is the full path to the %s runtime you wish to use [%s]? " % (app_name, app_path))
path = path if path else app_path
while not os.access(path, os.X_OK):
path = input(
" '%s' either does not exist or you don't have permission to run it, please enter the full path to '%s': " % (
path, app_name))
app_path = path
args = input(" What additional arguments would you like to pass to %s [%s]? " % (app_name, app_args))
if args:
app_args = args
queue = input(" What queue/partition should %s run on [%s]? " % (app_name, job_parms['queue']))
if queue:
job_parms['queue'] = queue
mem = input(" How much memory (GB) will %s require [%s]? " % (app_name, job_parms['mem_requested']))
if re.match("^[1-9][0-9]*$", mem):
job_parms['mem_requested'] = mem
cpus = input(" How many CPUs do you want %s to use [%s]? " % (app_name, job_parms['num_cpus']))
if re.match("^[1-9][0-9]*$", cpus):
job_parms['num_cpus'] = cpus
hours = input(" How many hours will %s take to run [%s]? " % (app_name, job_parms['walltime']))
if re.match("^[1-9][0-9]*$", hours):
job_parms['walltime'] = hours
return app_name, app_path, app_args, job_parms
def _get_aligners(queue, args):
aligner_list = []
bwa_path = ""
print(
"\nThis pipeline currently supports four aligners: BWA, Bowtie2, Novoalign, and SNAP.\nYou can also provide pre-aligned BAM files, and you can choose as many options as you want.")
# response = input("\nWould you like to run BWA samp/se [N]?* ")
# if re.match('^[Yy]', response):
# bwa_path = _get_application_path("bwa")
# bwa_sampe_settings = _get_advanced_settings("BWA-sampe", bwa_path, "",
# {'num_cpus': '4', 'mem_requested': '10', 'walltime': '36',
# 'queue': queue, 'args': args})
# aligner_list.append(bwa_sampe_settings)
# logging.info(bwa_sampe_settings)
response = input("\nWould you like to run BWA mem [Y]? ")
if not re.match('^[Nn]', response):
if not bwa_path:
bwa_path = _get_application_path("bwa")
bwa_mem_settings = _get_advanced_settings("BWA-mem", bwa_path, "",
{'num_cpus': '4', 'mem_requested': '10', 'walltime': '36',
'queue': queue, 'args': args})
aligner_list.append(bwa_mem_settings)
logging.info(bwa_mem_settings)
response = input("\nWould you like to run Bowtie2 [Y]? ")
if not re.match('^[Nn]', response):
bt2_path = _get_application_path("bowtie2")
bt2_settings = _get_advanced_settings("Bowtie2", bt2_path, "",
{'num_cpus': '4', 'mem_requested': '10', 'walltime': '36', 'queue': queue,
'args': args})
aligner_list.append(bt2_settings)
logging.info(bt2_settings)
response = input("\nWould you like to run Novoalign [N]? ")
if re.match('^[Yy]', response):
novo_path = _get_application_path("novoalign")
novo_settings = _get_advanced_settings("Novoalign", novo_path, "-r all",
{'num_cpus': '4', 'mem_requested': '10', 'walltime': '36',
'queue': queue, 'args': args})
aligner_list.append(novo_settings)
logging.info(novo_settings)
response = input("\nWould you like to run SNAP [N]? ")
if re.match('^[Yy]', response):
snap_path = _get_application_path("snap_aligner")
#Add -M option by default to make SNAP generate traditional CIGAR strings in the BAM output to be more compatible with SNP callers
snap_settings = _get_advanced_settings("SNAP", snap_path, "-M",
{'num_cpus': '4', 'mem_requested': '10', 'walltime': '36',
'queue': queue, 'args': args})
aligner_list.append(snap_settings)
logging.info(snap_settings)
return aligner_list
def _get_snpcallers(queue, args):
snpcaller_list = []
using_gatk = False
print(
"\nThis pipeline currently supports four SNP callers: GATK, SolSNP, VarScan, and SAMtools, and you can provide VCF files.\nYou can choose as many options as you want.")
response = input("\nWould you like to run GATK [Y]? ")
if not re.match('^[Nn]', response):
gatk_path = _get_java_path("GenomeAnalysisTK.jar")
gatk_settings = _get_advanced_settings("GATK", gatk_path, "-stand_call_conf 100 -stand_emit_conf 100 -ploidy 1",
{'num_cpus': '4', 'mem_requested': '10', 'walltime': '36',
'queue': queue, 'args': args})
snpcaller_list.append(gatk_settings)
logging.info(gatk_settings)
using_gatk = True
response = input("\nWould you like to run SolSNP [N]? ")
if re.match('^[Yy]', response):
solsnp_path = _get_java_path("SolSNP.jar")
solsnp_settings = _get_advanced_settings("SolSNP", solsnp_path, "",
{'num_cpus': '4', 'mem_requested': '10', 'walltime': '36',
'queue': queue, 'args': args})
snpcaller_list.append(solsnp_settings)
logging.info(solsnp_settings)
response = input("\nWould you like to run VarScan [Y]? ")
if not re.match('^[Nn]', response):
varscan_path = _get_java_path("VarScan.jar")
varscan_settings = _get_advanced_settings("VarScan", varscan_path, "",
{'num_cpus': '4', 'mem_requested': '10', 'walltime': '36',
'queue': queue, 'args': args})
snpcaller_list.append(varscan_settings)
logging.info(varscan_settings)
response = input("\nWould you like to run SAMtools [Y]? ")
if not re.match('^[Nn]', response):
samtools_path = _get_application_path("bcftools")
samtools_settings = _get_advanced_settings("SAMtools", samtools_path, "",
{'num_cpus': '4', 'mem_requested': '10', 'walltime': '36',
'queue': queue, 'args': args})
snpcaller_list.append(samtools_settings)
logging.info(samtools_settings)
return snpcaller_list, using_gatk
def _get_job_submitter():
job_submitter = "invalid"
queue = ""
args = ""
response = input(
"\nWhat system do you use for job management (PBS/TORQUE, SLURM, SGE/OGE, and 'none' are currently supported) [PBS]? ")
while job_submitter == "invalid":
if re.match('^(PBS|Torque|qsub)$', response, re.IGNORECASE) or response == "":
job_submitter = "PBS"
elif re.match('^(SLURM|sbatch)$', response, re.IGNORECASE):
job_submitter = "SLURM"
elif re.match('^(SGE|OGE)', response, re.IGNORECASE):
job_submitter = "SGE"
elif re.match('^none$', response, re.IGNORECASE):
job_submitter = "NONE"
else:
response = input(" %s is not a valid job management system, please enter another [PBS]? " % response)
if job_submitter != "NONE":
queue = input(
" Would you like to specify a queue/partition to use for all jobs (leave blank to use default queue) []? ")
args = input(" What additional arguments do you need to pass to the job management system []? ")
return job_submitter, queue, args
def _get_trimming_parameters(queue, args):
import pkg_resources
trim_path = _get_java_path("trimmomatic*.jar")
adapter_file = pkg_resources.resource_filename('nasp', 'illumina_adapters_all.fasta')
quality_string = ''
response = input(" What adapter file are you using for trimming [%s]? " % adapter_file)
if response:
adapter_file = response
response = input(" Would you also like to perform quality trimming [N]? ")
if re.match('^[Yy]', response):
quality_string = 'SLIDINGWINDOW:5:20'
response = input(" What quality trimming parameters do you want to use [%s]? " % quality_string)
if response:
quality_string = response
min_length = "80"
response = input(" What is the minimum length read to keep after trimming [%s]? " % min_length)
if response:
min_length = response
arg_string = "ILLUMINACLIP:%s:2:30:10 %s MINLEN:%s" % (adapter_file, quality_string, min_length)
return _get_advanced_settings("ReadTrimmer", trim_path, arg_string,
{'num_cpus': '4', 'mem_requested': '6', 'walltime': '24', 'queue': queue, 'args': args})
def _get_user_input(reference, output_folder):
import sys
configuration = {}
cwd = os.getcwd()
print("Welcome to NASP version %s." % nasp_version)
if not output_folder:
output_folder = input("\nWhere would you like output files to be written [nasp_results]? ")
if not output_folder:
output_folder = "nasp_results"
output_folder = _expand_path(output_folder)
if os.path.exists(output_folder):
response = input(
"\nOutput folder %s already exists!\nFiles in it may be overwritten!\nShould we continue anyway [N]? " % output_folder)
if not re.match('^[Yy]', response):
print("Operation cancelled!")
quit()
else:
os.makedirs(output_folder)
configuration["output_folder"] = output_folder
logfile = os.path.join(output_folder, "runlog.txt")
logging.basicConfig(level=logging.DEBUG,
format='%(asctime)s %(levelname)-8s %(message)s',
datefmt='%m/%d/%Y %H:%M:%S',
filename=logfile,
filemode='w')
logging.info("$PATH="+os.environ.get("PATH", ""))
logging.info("$PYTHONPATH="+os.environ.get("PYTHONPATH", ""))
if "LOADEDMODULES" in os.environ:
logging.info("LOADEDMODULES="+os.environ.get("LOADEDMODULES", ""))
if not reference:
reference = input("\nWhere is the reference fasta file you would like to use? ")
while not os.access(reference, os.R_OK) or not os.path.isfile(reference):
reference = input("\n'%s' either does not exist or you don't have permission to access it, please enter the path to the reference fasta file you would like to use: " % reference)
reference = _expand_path(reference)
configuration["reference"] = _create_file_tuple(reference)
logging.info("Reference = %s", configuration["reference"])
response = input(
"\nDo you want to check the reference for duplicated regions\nand skip SNPs that fall in those regions [Y]? ")
configuration["find_dups"] = "False" if re.match('^[Nn]', response) else "True"
logging.info("FindDups = %s", configuration["find_dups"])
(job_submitter, queue, args) = _get_job_submitter()
configuration["job_submitter"] = job_submitter
logging.info("JobSubmitter = %s", configuration["job_submitter"])
name_match = re.search('^.*/(.*)$',
output_folder) # Warning, not OS-independent! Should find a better way to do this.
configuration["run_name"] = name_match.group(
1) # Temporary: setting the run name to be whatever the the output folder is named. Should ask user.
logging.info("RunName = %s", configuration["run_name"])
samtools_path = _find_executable("samtools")
configuration["samtools"] = ("Samtools", samtools_path, "", {})
logging.info("Samtools = %s", configuration["samtools"])
run_path = os.path.abspath(os.path.dirname(sys.argv[0]))
configuration["index"] = ("Index", run_path, "",
{'name': 'nasp_index', 'num_cpus': '1', 'mem_requested': '2', 'walltime': '4',
'queue': queue, 'args': args})
logging.info("Index = %s", configuration["index"])
fasta_list = _get_external_fastas(cwd, reference)
configuration["assemblies"] = fasta_list
if configuration["find_dups"] or len(fasta_list) > 0:
nucmer_path = _get_application_path("nucmer")
nucmer_args = ""
if len(fasta_list) > 0:
deltafilter_path = _get_application_path("delta-filter")
deltafilter_args = ""
response = input(" Would you like to set advanced NUCmer settings [N]? ")
if re.match('^[Yy]', response):
nucmer_args = input(
" What additional arguments would you like to pass to 'nucmer' while importing external genomes? ")
deltafilter_args = input(
" What additional arguments would you like to pass to 'delta-filter' while importing external genomes? ")
configuration["assembly_importer"] = ("AssemblyImporter", deltafilter_path, deltafilter_args,
{'num_cpus': '1', 'mem_requested': '4', 'walltime': '4',
'queue': queue, 'args': args})
logging.info("AssemblyImporter = %s", configuration["assembly_importer"])
configuration["dup_finder"] = ("DupFinder", nucmer_path, nucmer_args,
{'num_cpus': '1', 'mem_requested': '4', 'walltime': '4', 'queue': queue,
'args': args})
logging.info("DupFinder = %s", configuration["dup_finder"])
read_list = _get_reads(cwd)
configuration["reads"] = read_list
if len(read_list) > 0:
response = input("\nWould you like to use Trimmomatic to trim your reads first [N]? ")
if re.match('^[Yy]', response):
configuration["trim_reads"] = "True"
configuration["read_trimmer"] = _get_trimming_parameters(queue, args)
logging.info("ReadTrimmer = %s", configuration["read_trimmer"])
logging.info("Getting Aligners...")
configuration["aligners"] = _get_aligners(queue, args)
else:
configuration["aligners"] = []
bam_list = _get_bams(cwd)
configuration["alignments"] = bam_list
if len(bam_list) > 0:
configuration["bam_index"] = ("BamIndex", run_path, "",
{'name': 'nasp_bamindex', 'num_cpus': '1', 'mem_requested': '2', 'walltime': '4',
'queue': queue, 'args': args})
logging.info("BamIndex = %s", configuration["bam_index"])
if len(read_list) > 0 or len(bam_list) > 0:
logging.info("Getting SNP Callers...")
(configuration["snpcallers"], using_gatk) = _get_snpcallers(queue, args)
if using_gatk:
picard_path = _get_java_path("picard.jar")
configuration["picard"] = ("Picard", picard_path, "", {})
logging.info("Picard = %s", configuration["picard"])
else:
configuration["snpcallers"] = []
vcf_list = _get_vcfs(cwd)
configuration["vcfs"] = vcf_list
if len(read_list) > 0 or len(bam_list) > 0 or len(vcf_list) > 0:
coverage_filter = input(
"\nThis pipeline can do filtering based on coverage.\nIf you do not want filtering based on coverage, enter 0.\nWhat is your minimum coverage threshold [10]? ")
if not coverage_filter:
coverage_filter = 10
configuration["coverage_filter"] = str(coverage_filter)
logging.info("CoverageFilter = %s", configuration["coverage_filter"])
proportion_filter = input(
"\nThis pipeline can do filtering based on the proportion of reads that match the call made by the SNP caller.\nIf you do not want filtering based on proportion, enter 0.\nWhat is the minimum acceptable proportion [0.9]? ")
if not proportion_filter:
proportion_filter = 0.9
configuration["proportion_filter"] = str(proportion_filter)
logging.info("ProportionFilter = %s", configuration["proportion_filter"])
matrix_settings = _get_advanced_settings("MatrixGenerator", gonasp_path(), "", {'name':'nasp_matrix', 'num_cpus':'8', 'mem_requested':'8', 'walltime':'48', 'queue':queue, 'args':args})
configuration["matrix_generator"] = matrix_settings
logging.info("MatrixGenerator = %s", configuration["matrix_generator"])
include_allref_pos = input("\nDo you want to create a master_masked matrix that includes all positions with low-quality positions that failed the coverage or proportion filter masked with an 'N' [N]? ")
if re.match('^[Yy]', include_allref_pos):
configuration["filter_matrix_format"] = "include_allref_pos"
logging.info("FilterMatrixFormat = %s", configuration["filter_matrix_format"])
return configuration
[docs]def gonasp_path():
import sys
import pkg_resources
if sys.maxsize > 2**32:
matrix_path = pkg_resources.resource_filename('nasp', 'nasptool_linux_64')
else:
matrix_path = pkg_resources.resource_filename('nasp', 'nasptool_linux_32')
if os.path.exists(matrix_path):
return matrix_path
else:
return "nasptool_linux_64"
# def guess_job_manager():
# try:
# import os
# import shutil
# # Order matters.
# # SLURM includes a simple qsub wrapper
# # TORQUE and SGE/OGE use variations of qsub
# if shutil.which('sbatch'):
# return 'SLURM'
# elif os.environ.get('SGE_ROOT') is not None:
# return 'SGE'
# elif shutil.which('qsub'):
# return 'PBS'
# else:
# return 'NONE'
# except ImportError:
# return 'NONE'
[docs]def main():
import sys
import nasp.dispatcher as dispatcher
import nasp.configuration_parser as configuration_parser
# This is hack to forward commands to gonasp
if len(sys.argv) > 1 and sys.argv[1] in ['help', 'duplicates', 'frankenfasta', 'matrix', 'export']:
if sys.argv[1] == 'help' or any(arg in ['-h', '-help', '--help'] for arg in sys.argv) or len(sys.argv) < 3:
print("Requesting nasp internal command usage message. Enter 'nasp --help' for the nasp pipeline usage message.")
print("Unlike the nasp pipeline, these commands will not be automatically submitted to your job manager")
import subprocess
subprocess.call([gonasp_path()] + sys.argv[1:])
return
commandline_args = _parse_args()
if commandline_args.config:
configuration = configuration_parser.parse_config(commandline_args.config)
output_folder = configuration["output_folder"]
if os.path.exists(output_folder):
response = input(
"\nOutput folder %s already exists!\nFiles in it may be overwritten!\nShould we continue anyway [N]? " % output_folder)
if not re.match('^[Yy]', response):
print("Operation cancelled!")
quit()
else:
os.makedirs(output_folder)
logfile = os.path.join(output_folder, "runlog.txt")
logging.basicConfig(level=logging.DEBUG,
format='%(asctime)s %(levelname)-8s %(message)s',
datefmt='%m/%d/%Y %H:%M:%S',
filename=logfile,
filemode='w')
else:
configuration = _get_user_input(commandline_args.reference_fasta, commandline_args.output_folder)
configuration_parser.write_config(configuration)
dispatcher.begin(configuration)
if __name__ == "__main__":
main()