def omgprepare(args):
"""
%prog omgprepare ploidy anchorsfile blastfile
Prepare to run Sankoff's OMG algorithm to get orthologs.
"""
from jcvi.formats.blast import cscore
from jcvi.formats.base import DictFile
p = OptionParser(omgprepare.__doc__)
p.add_option("--norbh", action="store_true",
help="Disable RBH hits [default: %default]")
p.add_option("--pctid", default=0, type="int",
help="Percent id cutoff for RBH hits [default: %default]")
p.add_option("--cscore", default=90, type="int",
help="C-score cutoff for RBH hits [default: %default]")
p.set_stripnames()
p.set_beds()
opts, args = p.parse_args(args)
if len(args) != 3:
sys.exit(not p.print_help())
ploidy, anchorfile, blastfile = args
norbh = opts.norbh
pctid = opts.pctid
cs = opts.cscore
qbed, sbed, qorder, sorder, is_self = check_beds(anchorfile, p, opts)
fp = open(ploidy)
genomeidx = dict((x.split()[0], i) for i, x in enumerate(fp))
fp.close()
ploidy = DictFile(ploidy)
geneinfo(qbed, qorder, genomeidx, ploidy)
geneinfo(sbed, sorder, genomeidx, ploidy)
pf = blastfile.rsplit(".", 1)[0]
cscorefile = pf + ".cscore"
cscore([blastfile, "-o", cscorefile, "--cutoff=0", "--pct"])
ac = AnchorFile(anchorfile)
pairs = set((a, b) for a, b, i in ac.iter_pairs())
logging.debug("Imported {0} pairs from `{1}`.".format(len(pairs), anchorfile))
weightsfile = pf + ".weights"
fp = open(cscorefile)
fw = open(weightsfile, "w")
npairs = 0
for row in fp:
a, b, c, pct = row.split()
c, pct = float(c), float(pct)
c = int(c * 100)
if (a, b) not in pairs:
if norbh:
continue
if c < cs:
continue
if pct < pctid:
continue
c /= 10 # This severely penalizes RBH against synteny
print >> fw, "\t".join((a, b, str(c)))
npairs += 1
fw.close()
logging.debug("Write {0} pairs to `{1}`.".format(npairs, weightsfile))
def ortholog(args):
"""
%prog ortholog species_a species_b
Run a sensitive pipeline to find orthologs between two species a and b.
The pipeline runs LAST and generate .lifted.anchors.
`--full` mode would assume 1-to-1 quota synteny blocks as the backbone of
such predictions. Extra orthologs will be recruited from reciprocal best
match (RBH).
"""
from jcvi.apps.align import last as last_main
from jcvi.compara.blastfilter import main as blastfilter_main
from jcvi.compara.quota import main as quota_main
from jcvi.compara.synteny import scan, mcscan, liftover
from jcvi.formats.blast import cscore, filter
p = OptionParser(ortholog.__doc__)
p.add_option("--full", default=False, action="store_true",
help="Run in full mode, including blocks and RBH")
p.add_option("--cscore", default=0.7, type="float",
help="C-score cutoff [default: %default]")
p.add_option("--dist", default=20, type="int",
help="Extent of flanking regions to search")
p.add_option("--quota", help="Quota align parameter")
opts, args = p.parse_args(args)
if len(args) != 2:
sys.exit(not p.print_help())
a, b = args
abed, afasta = a + ".bed", a + ".cds"
bbed, bfasta = b + ".bed", b + ".cds"
ccscore = opts.cscore
quota = opts.quota
dist = "--dist={0}".format(opts.dist)
aprefix = afasta.split(".")[0]
bprefix = bfasta.split(".")[0]
pprefix = ".".join((aprefix, bprefix))
qprefix = ".".join((bprefix, aprefix))
last = pprefix + ".last"
if need_update((afasta, bfasta), last):
last_main([bfasta, afasta])
if a == b:
last = filter([last, "--hitlen=0", "--pctid=98", "--inverse", "--noself"])
filtered_last = last + ".filtered"
if need_update(last, filtered_last):
blastfilter_main([last, "--cscore={0}".format(ccscore)])
anchors = pprefix + ".anchors"
lifted_anchors = pprefix + ".lifted.anchors"
if not opts.full:
if need_update(filtered_last, lifted_anchors):
scan([filtered_last, anchors, dist,
"--liftover={0}".format(last)])
if quota:
quota_main([lifted_anchors,
"--quota={0}".format(quota), "--screen"])
return
if need_update(filtered_last, anchors):
scan([filtered_last, anchors, dist])
ooanchors = pprefix + ".1x1.anchors"
if need_update(anchors, ooanchors):
quota_main([anchors, "--quota=1:1", "--screen"])
lifted_anchors = pprefix + ".1x1.lifted.anchors"
if need_update((last, ooanchors), lifted_anchors):
liftover([last, ooanchors, dist])
pblocks = pprefix + ".1x1.blocks"
qblocks = qprefix + ".1x1.blocks"
if need_update(lifted_anchors, [pblocks, qblocks]):
mcscan([abed, lifted_anchors, "--iter=1", "-o", pblocks])
mcscan([bbed, lifted_anchors, "--iter=1", "-o", qblocks])
rbh = pprefix + ".rbh"
if need_update(last, rbh):
cscore([last, "-o", rbh])
portho = pprefix + ".ortholog"
qortho = qprefix + ".ortholog"
if need_update([pblocks, qblocks, rbh], [portho, qortho]):
make_ortholog(pblocks, rbh, portho)
make_ortholog(qblocks, rbh, qortho)
def ortholog(args):
"""
%prog ortholog species_a species_b
Run a sensitive pipeline to find orthologs between two species a and b.
The pipeline runs LAST and generate .lifted.anchors.
`--full` mode would assume 1-to-1 quota synteny blocks as the backbone of
such predictions. Extra orthologs will be recruited from reciprocal best
match (RBH).
"""
from jcvi.apps.last import main as last_main
from jcvi.compara.blastfilter import main as blastfilter_main
from jcvi.compara.quota import main as quota_main
from jcvi.compara.synteny import scan, mcscan, liftover
from jcvi.formats.blast import cscore, filter
p = OptionParser(ortholog.__doc__)
p.add_option("--full",
default=False,
action="store_true",
help="Run in full mode, including blocks and RBH")
p.add_option("--cscore",
default=0.7,
type="float",
help="C-score cutoff [default: %default]")
p.add_option("--dist",
default=20,
type="int",
help="Extent of flanking regions to search")
opts, args = p.parse_args(args)
if len(args) != 2:
sys.exit(not p.print_help())
a, b = args
abed, afasta = a + ".bed", a + ".cds"
bbed, bfasta = b + ".bed", b + ".cds"
ccscore = opts.cscore
dist = "--dist={0}".format(opts.dist)
aprefix = afasta.split(".")[0]
bprefix = bfasta.split(".")[0]
pprefix = ".".join((aprefix, bprefix))
qprefix = ".".join((bprefix, aprefix))
last = pprefix + ".last"
if need_update((afasta, bfasta), last):
last_main([bfasta, afasta, "-o", last])
if a == b:
last = filter([last, "--hitlen=0", "--pctid=98", "--inverse"])
filtered_last = last + ".filtered"
if need_update(last, filtered_last):
blastfilter_main([last, "--cscore={0}".format(ccscore)])
anchors = pprefix + ".anchors"
lifted_anchors = pprefix + ".lifted.anchors"
if not opts.full:
if need_update(filtered_last, lifted_anchors):
scan([filtered_last, anchors, dist, "--liftover={0}".format(last)])
return
if need_update(filtered_last, anchors):
scan([filtered_last, anchors, dist])
ooanchors = pprefix + ".1x1.anchors"
if need_update(anchors, ooanchors):
quota_main([anchors, "--quota=1:1", "--screen"])
lifted_anchors = pprefix + ".1x1.lifted.anchors"
if need_update((last, ooanchors), lifted_anchors):
liftover([last, ooanchors, dist])
pblocks = pprefix + ".1x1.blocks"
qblocks = qprefix + ".1x1.blocks"
if need_update(lifted_anchors, [pblocks, qblocks]):
mcscan([abed, lifted_anchors, "--iter=1", "-o", pblocks])
mcscan([bbed, lifted_anchors, "--iter=1", "-o", qblocks])
rbh = pprefix + ".rbh"
if need_update(last, rbh):
cscore([last, "-o", rbh])
portho = pprefix + ".ortholog"
qortho = qprefix + ".ortholog"
if need_update([pblocks, qblocks, rbh], [portho, qortho]):
make_ortholog(pblocks, rbh, portho)
make_ortholog(qblocks, rbh, qortho)
def omgprepare(args):
"""
%prog omgprepare ploidy anchorsfile blastfile
Prepare to run Sankoff's OMG algorithm to get orthologs.
"""
from jcvi.formats.blast import cscore
from jcvi.formats.base import DictFile
p = OptionParser(omgprepare.__doc__)
p.add_option("--norbh",
action="store_true",
help="Disable RBH hits [default: %default]")
p.add_option("--pctid",
default=0,
type="int",
help="Percent id cutoff for RBH hits [default: %default]")
p.add_option("--cscore",
default=90,
type="int",
help="C-score cutoff for RBH hits [default: %default]")
p.set_stripnames()
p.set_beds()
opts, args = p.parse_args(args)
if len(args) != 3:
sys.exit(not p.print_help())
ploidy, anchorfile, blastfile = args
norbh = opts.norbh
pctid = opts.pctid
cs = opts.cscore
qbed, sbed, qorder, sorder, is_self = check_beds(anchorfile, p, opts)
fp = open(ploidy)
genomeidx = dict((x.split()[0], i) for i, x in enumerate(fp))
fp.close()
ploidy = DictFile(ploidy)
geneinfo(qbed, qorder, genomeidx, ploidy)
geneinfo(sbed, sorder, genomeidx, ploidy)
pf = blastfile.rsplit(".", 1)[0]
cscorefile = pf + ".cscore"
cscore([blastfile, "-o", cscorefile, "--cutoff=0", "--pct"])
ac = AnchorFile(anchorfile)
pairs = set((a, b) for a, b, i in ac.iter_pairs())
logging.debug("Imported {0} pairs from `{1}`.".format(
len(pairs), anchorfile))
weightsfile = pf + ".weights"
fp = open(cscorefile)
fw = open(weightsfile, "w")
npairs = 0
for row in fp:
a, b, c, pct = row.split()
c, pct = float(c), float(pct)
c = int(c * 100)
if (a, b) not in pairs:
if norbh:
continue
if c < cs:
continue
if pct < pctid:
continue
c /= 10 # This severely penalizes RBH against synteny
print >> fw, "\t".join((a, b, str(c)))
npairs += 1
fw.close()
logging.debug("Write {0} pairs to `{1}`.".format(npairs, weightsfile))
def ortholog(args):
"""
%prog ortholog species_a species_b
Run a sensitive pipeline to find orthologs between two species a and b.
The pipeline runs LAST and generate .lifted.anchors.
`--full` mode would assume 1-to-1 quota synteny blocks as the backbone of
such predictions. Extra orthologs will be recruited from reciprocal best
match (RBH).
"""
from jcvi.apps.align import last as last_main
from jcvi.compara.blastfilter import main as blastfilter_main
from jcvi.compara.quota import main as quota_main
from jcvi.compara.synteny import scan, mcscan, liftover
from jcvi.formats.blast import cscore, filter
p = OptionParser(ortholog.__doc__)
p.add_option(
"--dbtype",
default="nucl",
choices=("nucl", "prot"),
help="Molecule type of subject database",
)
p.add_option(
"--full",
default=False,
action="store_true",
help="Run in full 1x1 mode, including blocks and RBH",
)
p.add_option("--cscore", default=0.7, type="float", help="C-score cutoff")
p.add_option(
"--dist", default=20, type="int", help="Extent of flanking regions to search"
)
p.add_option(
"-n",
"--min_size",
dest="n",
type="int",
default=4,
help="minimum number of anchors in a cluster",
)
p.add_option("--quota", help="Quota align parameter")
p.add_option(
"--no_strip_names",
default=False,
action="store_true",
help="Do not strip alternative splicing (e.g. At5g06540.1 -> At5g06540)",
)
p.set_cpus()
p.set_dotplot_opts()
opts, args = p.parse_args(args)
if len(args) != 2:
sys.exit(not p.print_help())
a, b = args
dbtype = opts.dbtype
suffix = ".cds" if dbtype == "nucl" else ".pep"
abed, afasta = a + ".bed", a + suffix
bbed, bfasta = b + ".bed", b + suffix
ccscore = opts.cscore
quota = opts.quota
dist = "--dist={0}".format(opts.dist)
minsize_flag = "--min_size={}".format(opts.n)
cpus_flag = "--cpus={}".format(opts.cpus)
aprefix = afasta.split(".")[0]
bprefix = bfasta.split(".")[0]
pprefix = ".".join((aprefix, bprefix))
qprefix = ".".join((bprefix, aprefix))
last = pprefix + ".last"
if need_update((afasta, bfasta), last):
last_main([bfasta, afasta, cpus_flag], dbtype)
if a == b:
lastself = last + ".P98L0.inverse"
if need_update(last, lastself):
filter([last, "--hitlen=0", "--pctid=98", "--inverse", "--noself"])
last = lastself
filtered_last = last + ".filtered"
if need_update(last, filtered_last):
if opts.no_strip_names:
blastfilter_main([last, "--cscore={0}".format(ccscore), "--no_strip_names"])
else:
blastfilter_main([last, "--cscore={0}".format(ccscore)])
anchors = pprefix + ".anchors"
lifted_anchors = pprefix + ".lifted.anchors"
pdf = pprefix + ".pdf"
if not opts.full:
if need_update(filtered_last, lifted_anchors):
dargs = [
filtered_last,
anchors,
minsize_flag,
dist,
"--liftover={0}".format(last),
]
if opts.no_strip_names:
dargs += [
"--no_strip_names",
]
scan(dargs)
if quota:
quota_main([lifted_anchors, "--quota={0}".format(quota), "--screen"])
if need_update(anchors, pdf):
from jcvi.graphics.dotplot import dotplot_main
dargs = [anchors]
if opts.nostdpf:
dargs += ["--nostdpf"]
if opts.nochpf:
dargs += ["--nochpf"]
if opts.skipempty:
dargs += ["--skipempty"]
if opts.genomenames:
dargs += ["--genomenames", opts.genomenames]
if opts.theme:
dargs += ["--theme", opts.theme]
dotplot_main(dargs)
return
if need_update(filtered_last, anchors):
if opts.no_strip_names:
scan([filtered_last, anchors, dist, "--no_strip_names"])
else:
scan([filtered_last, anchors, dist])
ooanchors = pprefix + ".1x1.anchors"
if need_update(anchors, ooanchors):
quota_main([anchors, "--quota=1:1", "--screen"])
lifted_anchors = pprefix + ".1x1.lifted.anchors"
if need_update((last, ooanchors), lifted_anchors):
if opts.no_strip_names:
liftover([last, ooanchors, dist, "--no_strip_names"])
else:
liftover([last, ooanchors, dist])
pblocks = pprefix + ".1x1.blocks"
qblocks = qprefix + ".1x1.blocks"
if need_update(lifted_anchors, [pblocks, qblocks]):
mcscan([abed, lifted_anchors, "--iter=1", "-o", pblocks])
mcscan([bbed, lifted_anchors, "--iter=1", "-o", qblocks])
rbh = pprefix + ".rbh"
if need_update(last, rbh):
cscore([last, "-o", rbh])
portho = pprefix + ".ortholog"
qortho = qprefix + ".ortholog"
if need_update([pblocks, qblocks, rbh], [portho, qortho]):
make_ortholog(pblocks, rbh, portho)
make_ortholog(qblocks, rbh, qortho)
def ortholog(args):
"""
%prog ortholog a.bed a.cds b.bed b.cds
Run a sensitive pipeline to find orthologs between two species a and b.
The pipeline runs LAST and 1-to-1 quota synteny blocks as the backbone of
such predictions. Extra orthologs will be recruited from reciprocal best
match (RBH).
"""
from jcvi.apps.last import main as last_main
from jcvi.compara.blastfilter import main as blastfilter_main
from jcvi.compara.quota import main as quota_main
from jcvi.compara.synteny import scan, mcscan, liftover
from jcvi.formats.blast import cscore
p = OptionParser(ortholog.__doc__)
p.add_option("--cscore", default=0.99, type="float", help="C-score cutoff [default: %default]")
opts, args = p.parse_args(args)
if len(args) != 4:
sys.exit(not p.print_help())
abed, afasta, bbed, bfasta = args
ccscore = opts.cscore
aprefix = afasta.split(".")[0]
bprefix = bfasta.split(".")[0]
pprefix = ".".join((aprefix, bprefix))
qprefix = ".".join((bprefix, aprefix))
last = pprefix + ".last"
if need_update((afasta, bfasta), last):
last_main([bfasta, afasta, "-o", last])
filtered_last = last + ".filtered"
bstring = ["--qbed=" + abed, "--sbed=" + bbed]
if need_update(last, filtered_last):
blastfilter_main([last, "--cscore={0}".format(ccscore), "--tandem_Nmax=10"] + bstring)
anchors = pprefix + ".anchors"
lifted_anchors = pprefix + ".lifted.anchors"
if need_update(filtered_last, lifted_anchors):
scan([filtered_last, anchors] + bstring)
ooanchors = pprefix + ".1x1.anchors"
if need_update(anchors, ooanchors):
quota_main([anchors, "--quota=1:1", "--screen"] + bstring)
lifted_anchors = pprefix + ".1x1.lifted.anchors"
if need_update((last, ooanchors), lifted_anchors):
liftover([last, ooanchors] + bstring)
pblocks = pprefix + ".1x1.blocks"
qblocks = qprefix + ".1x1.blocks"
if need_update(lifted_anchors, [pblocks, qblocks]):
mcscan([abed, lifted_anchors, "--iter=1", "-o", pblocks])
mcscan([bbed, lifted_anchors, "--iter=1", "-o", qblocks])
rbh = pprefix + ".rbh"
if need_update(last, rbh):
cscore([last, "-o", rbh])
portho = pprefix + ".ortholog"
qortho = qprefix + ".ortholog"
if need_update([pblocks, qblocks, rbh], [portho, qortho]):
make_ortholog(pblocks, rbh, portho)
make_ortholog(qblocks, rbh, qortho)
def ortholog(args):
"""
%prog ortholog species_a species_b
Run a sensitive pipeline to find orthologs between two species a and b.
The pipeline runs LAST and generate .lifted.anchors.
`--full` mode would assume 1-to-1 quota synteny blocks as the backbone of
such predictions. Extra orthologs will be recruited from reciprocal best
match (RBH).
"""
from jcvi.apps.align import last as last_main
from jcvi.compara.blastfilter import main as blastfilter_main
from jcvi.compara.quota import main as quota_main
from jcvi.compara.synteny import scan, mcscan, liftover
from jcvi.formats.blast import cscore, filter
p = OptionParser(ortholog.__doc__)
p.add_option("--dbtype", default="nucl",
choices=("nucl", "prot"),
help="Molecule type of subject database")
p.add_option("--full", default=False, action="store_true",
help="Run in full mode, including blocks and RBH")
p.add_option("--cscore", default=0.7, type="float",
help="C-score cutoff [default: %default]")
p.add_option("--dist", default=20, type="int",
help="Extent of flanking regions to search")
p.add_option("--quota", help="Quota align parameter")
p.add_option("--nostdpf", default=False, action="store_true",
help="Do not standardize contig names")
p.add_option("--no_strip_names", default=False, action="store_true",
help="Do not strip alternative splicing "
"(e.g. At5g06540.1 -> At5g06540)")
opts, args = p.parse_args(args)
if len(args) != 2:
sys.exit(not p.print_help())
a, b = args
dbtype = opts.dbtype
suffix = ".cds" if dbtype == "nucl" else ".pep"
abed, afasta = a + ".bed", a + suffix
bbed, bfasta = b + ".bed", b + suffix
ccscore = opts.cscore
quota = opts.quota
dist = "--dist={0}".format(opts.dist)
aprefix = afasta.split(".")[0]
bprefix = bfasta.split(".")[0]
pprefix = ".".join((aprefix, bprefix))
qprefix = ".".join((bprefix, aprefix))
last = pprefix + ".last"
if need_update((afasta, bfasta), last):
last_main([bfasta, afasta], dbtype)
if a == b:
lastself = last + ".P98L0.inverse"
if need_update(last, lastself):
filter([last, "--hitlen=0", "--pctid=98", "--inverse", "--noself"])
last = lastself
filtered_last = last + ".filtered"
if need_update(last, filtered_last):
if opts.no_strip_names:
blastfilter_main([last, "--cscore={0}".format(ccscore), "--no_strip_names"])
else:
blastfilter_main([last, "--cscore={0}".format(ccscore)])
anchors = pprefix + ".anchors"
lifted_anchors = pprefix + ".lifted.anchors"
pdf = pprefix + ".pdf"
if not opts.full:
if need_update(filtered_last, lifted_anchors):
if opts.no_strip_names:
scan([filtered_last, anchors, dist,
"--liftover={0}".format(last), "--no_strip_names"])
else:
scan([filtered_last, anchors, dist,
"--liftover={0}".format(last)])
if quota:
quota_main([lifted_anchors,
"--quota={0}".format(quota), "--screen"])
if need_update(anchors, pdf):
from jcvi.graphics.dotplot import dotplot_main
dargs = [anchors]
if opts.nostdpf:
dargs += ["--nostdpf", "--skipempty"]
dotplot_main(dargs)
return
if need_update(filtered_last, anchors):
if opts.no_strip_names:
scan([filtered_last, anchors, dist, "--no_strip_names"])
else:
scan([filtered_last, anchors, dist])
ooanchors = pprefix + ".1x1.anchors"
if need_update(anchors, ooanchors):
quota_main([anchors, "--quota=1:1", "--screen"])
lifted_anchors = pprefix + ".1x1.lifted.anchors"
if need_update((last, ooanchors), lifted_anchors):
if opts.no_strip_names:
liftover([last, ooanchors, dist, "--no_strip_names"])
else:
liftover([last, ooanchors, dist])
pblocks = pprefix + ".1x1.blocks"
qblocks = qprefix + ".1x1.blocks"
if need_update(lifted_anchors, [pblocks, qblocks]):
mcscan([abed, lifted_anchors, "--iter=1", "-o", pblocks])
mcscan([bbed, lifted_anchors, "--iter=1", "-o", qblocks])
rbh = pprefix + ".rbh"
if need_update(last, rbh):
cscore([last, "-o", rbh])
portho = pprefix + ".ortholog"
qortho = qprefix + ".ortholog"
if need_update([pblocks, qblocks, rbh], [portho, qortho]):
make_ortholog(pblocks, rbh, portho)
make_ortholog(qblocks, rbh, qortho)