def computeFit(options):
if options.tree is not None:
tree = options.tree
else:
tree = getHalTree(options.hal)
cmd = "phyloFit --tree \"%s\" --subst-mod %s --sym-freqs %s --out-root %s" % (
tree, options.substMod, options.outMafSS,
os.path.splitext(options.outMod)[0])
if options.error is not None:
cmd += " --error %s " % options.error
runShellCommand(cmd)
runShellCommand("rm -f %s" % options.outMafSS)
def main(argv=None):
if argv is None:
argv = sys.argv
parser = argparse.ArgumentParser(
formatter_class=argparse.ArgumentDefaultsHelpFormatter,
description="Compute a neutral substitution model for use with "
"phyloP or halPhlyoP")
parser.add_argument("hal", help="input hal")
parser.add_argument("refGenome", help="Name of reference genome")
parser.add_argument("bedDir", help="BED file or directory containing BED "
"files. By "
"default, these files are interpreted to contain only"
" coordinates of coding exons, and fourfold degenerate"
" sites will automatically be extracted from them."
" To disable this behaviour and train on the entire "
" file, use the --no4d option.", default=None)
parser.add_argument("outMod", help="Path to output model file")
parser.add_argument("--no4d", help="Do not extract fourfold degenerate"
" positions from the input bed files. Rather use "
"all bases they contain.",
default=False, action="store_true")
parser.add_argument("--numProc",
help="Maximum number of processes for hal2maf.",
type=int, default=1)
parser.add_argument("--noAncestors",
help="Don't write ancestral genomes in hal2maf",
action="store_true", default=False)
parser.add_argument("--maxBedLines",
help="Split bed files so they have at most this many"
" lines",
type=int, default=None)
parser.add_argument("--sliceSize",
help="Slice size for hal2maf.",
type=int, default=None)
parser.add_argument("--tree",
help="String describing phylogeny in NEWICK format "
"that will be used instead of the tree stored in the"
" HAL file. This tree should contain all the species"
" in the alignment. Note that it is best to enclose"
" this string in quotes",
default=None)
parser.add_argument("--targetGenomes", default=None, nargs='+',
help="space separated list of targetGenomes to pass to "
"hal2maf. If used, the tree given to --tree should match.")
parser.add_argument("--substMod", help="Substitution model for phyloFit"
": valid options are JC69|F81|HKY85|HKY85+Gap|REV|"
"SSREV|UNREST|R2|R2S|U2|U2S|R3|R3S|U3|U3S",
default = "SSREV")
parser.add_argument("--noModFreqs", help="By default, equilibrium "
"frequencies for the nucleotides of the trained model"
" are corrected with the observed frequencies of "
"the reference genome (using the PHAST modFreqs"
" tool. This flag disables this step, and keeps the"
" trained frequencies", action="store_true",
default=False)
parser.add_argument("--error", help="File in which to output confidence"
" intervals for the parameters in the model",
default=None)
args = parser.parse_args()
# validate inputs
if not os.path.isfile(args.hal):
raise RuntimeError("Input hal file %s not found" % args.hal)
if not os.path.exists(args.bedDir):
raise RuntimeError("%s not found" % args.bedDir)
# validarte substitution model
if not args.substMod in "JC69|F81|HKY85|HKY85+Gap|REV|SSREV|UNREST|R2|R2S|U2|U2S|R3|R3S|U3|U3S".split("|"):
raise RuntimeError("Invalid substitution model: %s" % args.substMod)
# validate BEDs
if os.path.isdir(args.bedDir):
args.bedFiles = [os.path.join(args.bedDir, f) for f
in os.listdir(args.bedDir)
if os.path.isfile(os.path.join(args.bedDir, f))]
else:
args.bedFiles = [args.bedDir]
# test output is writeable and has valid extension
outTest = open(args.outMod, "w")
if not outTest:
raise RuntimeError("Unable to open output %s" % args.outMod)
if os.path.splitext(args.outMod)[1] != ".mod":
raise RuntimeError("Output model must have .mod extension")
# if targetGenomes is set, use those. Otherwise, extract from HAL
if args.targetGenomes is not None:
args.halGenomes = args.targetGenomes
else:
args.halGenomes = getHalGenomes(args.hal)
# if tree is set, use that. Otherwise, extract from HAL
if args.tree is None:
args.tree = getHalTree(args.hal)
# Make sure that all members of halGenomes and tree are in the actual HAL
halTree = getHalTree(args.hal)
if args.refGenome not in halTree:
raise RuntimeError("Reference genome %s not found." % args.refGenome)
for targetGenome in args.halGenomes:
if targetGenome not in halTree:
raise RuntimeError("Target genome %s not in HAL." % targetGenome)
if targetGenome not in args.tree:
raise RuntimeError("Target genome %s not in --tree." % targetGenome)
args.halGenomes = ','.join(args.halGenomes)
args.outDir = os.path.dirname(args.outMod)
args.outName = os.path.splitext(os.path.basename(args.outMod))[0]
args.outMafName = args.outName + "_halPhyloPTrain_temp.maf"
args.outMafPath = os.path.join(args.outDir, args.outMafName)
args.outMafAllPaths = args.outMafPath.replace("_halPhyloPTrain_temp.maf",
"_halPhyloPTrain_temp*.maf")
args.outMafSS = args.outMafPath.replace("_halPhyloPTrain_temp.maf",
"_halPhyloPTrain_temp.ss")
computeModel(args)
def main(argv=None):
if argv is None:
argv = sys.argv
parser = argparse.ArgumentParser(
formatter_class=argparse.ArgumentDefaultsHelpFormatter,
description="Compute a neutral substitution model for use with "
"phyloP or halPhlyoP")
parser.add_argument("hal", help="input hal")
parser.add_argument("refGenome", help="Name of reference genome")
parser.add_argument("bedDir",
help="BED file or directory containing BED "
"files. By "
"default, these files are interpreted to contain only"
" coordinates of coding exons, and fourfold degenerate"
" sites will automatically be extracted from them."
" To disable this behaviour and train on the entire "
" file, use the --no4d option.",
default=None)
parser.add_argument("outMod", help="Path to output model file")
parser.add_argument("--no4d",
help="Do not extract fourfold degenerate"
" positions from the input bed files. Rather use "
"all bases they contain.",
default=False,
action="store_true")
parser.add_argument("--numProc",
help="Maximum number of processes for hal2maf.",
type=int,
default=1)
parser.add_argument("--noAncestors",
help="Don't write ancestral genomes in hal2maf",
action="store_true",
default=False)
parser.add_argument("--maxBedLines",
help="Split bed files so they have at most this many"
" lines",
type=int,
default=None)
parser.add_argument("--tree",
help="String describing phylogeny in NEWICK format "
"that will be used instead of the tree stored in the"
" HAL file. This tree should contain all the species"
" in the alignment. Note that it is best to enclose"
" this string in quotes",
default=None)
parser.add_argument(
"--targetGenomes",
default=None,
nargs='+',
help="space separated list of targetGenomes to pass to "
"hal2maf. If used, the tree given to --tree should match.")
parser.add_argument("--substMod",
help="Substitution model for phyloFit"
": valid options are JC69|F81|HKY85|HKY85+Gap|REV|"
"SSREV|UNREST|R2|R2S|U2|U2S|R3|R3S|U3|U3S",
default="SSREV")
parser.add_argument("--noModFreqs",
help="By default, equilibrium "
"frequencies for the nucleotides of the trained model"
" are corrected with the observed frequencies of "
"the reference genome (using the PHAST modFreqs"
" tool. This flag disables this step, and keeps the"
" trained frequencies",
action="store_true",
default=False)
parser.add_argument("--precision",
help="Precision to pass to phyloFit (default MED)",
choices=["HIGH", "MED", "LOW"],
default="MED")
parser.add_argument("--error",
help="File in which to output confidence"
" intervals for the parameters in the model",
default=None)
args = parser.parse_args()
# validate inputs
if not os.path.isfile(args.hal):
raise RuntimeError("Input hal file %s not found" % args.hal)
if not os.path.exists(args.bedDir):
raise RuntimeError("%s not found" % args.bedDir)
# validarte substitution model
if not args.substMod in "JC69|F81|HKY85|HKY85+Gap|REV|SSREV|UNREST|R2|R2S|U2|U2S|R3|R3S|U3|U3S".split(
"|"):
raise RuntimeError("Invalid substitution model: %s" % args.substMod)
# validate BEDs
if os.path.isdir(args.bedDir):
args.bedFiles = [
os.path.join(args.bedDir, f) for f in os.listdir(args.bedDir)
if os.path.isfile(os.path.join(args.bedDir, f))
]
else:
args.bedFiles = [args.bedDir]
# test output is writeable and has valid extension
outTest = open(args.outMod, "w")
if not outTest:
raise RuntimeError("Unable to open output %s" % args.outMod)
if os.path.splitext(args.outMod)[1] != ".mod":
raise RuntimeError("Output model must have .mod extension")
# if targetGenomes is set, use those. Otherwise, extract from HAL
if args.targetGenomes is not None:
args.halGenomes = args.targetGenomes
else:
args.halGenomes = getHalGenomes(args.hal)
# if tree is set, use that. Otherwise, extract from HAL
if args.tree is None:
args.tree = getHalTree(args.hal)
# Make sure that all members of halGenomes and tree are in the actual HAL
halTree = getHalTree(args.hal)
if args.refGenome not in halTree:
raise RuntimeError("Reference genome %s not found." % args.refGenome)
for targetGenome in args.halGenomes:
if targetGenome not in halTree:
raise RuntimeError("Target genome %s not in HAL." % targetGenome)
if targetGenome not in args.tree:
raise RuntimeError("Target genome %s not in --tree." %
targetGenome)
args.halGenomes = ','.join(args.halGenomes)
args.outDir = os.path.dirname(args.outMod)
args.outName = os.path.splitext(os.path.basename(args.outMod))[0]
# Random suffix so two runs don't collide
suffix = "".join(
[random.choice(string.ascii_uppercase) for _ in xrange(7)])
args.outMafName = args.outName + "_halPhyloPTrain_temp_%s.maf" % suffix
args.outMafPath = os.path.join(args.outDir, args.outMafName)
args.outMafAllPaths = args.outMafPath.replace(
"_halPhyloPTrain_temp_%s.maf" % suffix,
"_halPhyloPTrain_temp_%s*.maf" % suffix)
# replace .maf suffix with .ss
args.outMafSS = args.outMafPath[:-4] + ".ss"
computeModel(args)
