def get_accuracy(query, gff_file, evidences_bed, sizesfile, type, key):
"""
Get sensitivity, specificity and accuracy given gff_file, and a query range
that look like "chr1:1-10000".
"""
from jcvi.formats.bed import evaluate
bed_file = get_bed_file(gff_file, type, key)
b = evaluate([bed_file, evidences_bed, sizesfile, "--query={0}".format(query)])
return b
def get_accuracy(query, gff_file, evidences_bed, sizesfile, type, key):
"""
Get sensitivity, specificity and accuracy given gff_file, and a query range
that look like "chr1:1-10000".
"""
from jcvi.formats.bed import evaluate
bed_file = get_bed_file(gff_file, type, key)
b = evaluate([bed_file, evidences_bed, sizesfile, "--query={0}".format(query)])
return b
def batcheval(args):
"""
%prog batcheval model.ids gff_file evidences.bed fastafile
Get the accuracy for a list of models against evidences in the range of the
genes. For example:
$ %prog batcheval all.gff3 isoforms.ids proteins.bed scaffolds.fasta
Outfile contains the scores for the models can be found in models.scores
"""
from jcvi.formats.bed import evaluate
from jcvi.formats.gff import make_index
p = OptionParser(evaluate.__doc__)
p.add_option(
"--type",
default="CDS",
help="list of features to extract, use comma to separate (e.g."
"'five_prime_UTR,CDS,three_prime_UTR') [default: %default]")
opts, args = p.parse_args(args)
if len(args) != 4:
sys.exit(not p.print_help())
model_ids, gff_file, evidences_bed, fastafile = args
type = set(opts.type.split(","))
g = make_index(gff_file)
fp = open(model_ids)
prefix = model_ids.rsplit(".", 1)[0]
fwscores = open(prefix + ".scores", "w")
for row in fp:
cid = row.strip()
b = next(g.parents(cid, 1))
query = "{0}:{1}-{2}".format(b.chrom, b.start, b.stop)
children = [c for c in g.children(cid, 1)]
cidbed = prefix + ".bed"
fw = open(cidbed, "w")
for c in children:
if c.featuretype not in type:
continue
fw.write(c.to_bed())
fw.close()
b = evaluate(
[cidbed, evidences_bed, fastafile, "--query={0}".format(query)])
print("\t".join((cid, b.score)), file=fwscores)
fwscores.flush()
def batcheval(args):
"""
%prog batcheval model.ids gff_file evidences.bed fastafile
Get the accuracy for a list of models against evidences in the range of the
genes. For example:
$ %prog batcheval all.gff3 isoforms.ids proteins.bed scaffolds.fasta
Outfile contains the scores for the models can be found in models.scores
"""
from jcvi.formats.bed import evaluate
from jcvi.formats.gff import make_index
p = OptionParser(evaluate.__doc__)
p.add_option("--type", default="CDS",
help="list of features to extract, use comma to separate (e.g."
"'five_prime_UTR,CDS,three_prime_UTR') [default: %default]")
opts, args = p.parse_args(args)
if len(args) != 4:
sys.exit(not p.print_help())
model_ids, gff_file, evidences_bed, fastafile = args
type = set(opts.type.split(","))
g = make_index(gff_file)
fp = open(model_ids)
prefix = model_ids.rsplit(".", 1)[0]
fwscores = open(prefix + ".scores", "w")
for row in fp:
cid = row.strip()
b = g.parents(cid, 1).next()
query = "{0}:{1}-{2}".format(b.chrom, b.start, b.stop)
children = [c for c in g.children(cid, 1)]
cidbed = prefix + ".bed"
fw = open(cidbed, "w")
for c in children:
if c.featuretype not in type:
continue
fw.write(c.to_bed())
fw.close()
b = evaluate([cidbed, evidences_bed, fastafile, "--query={0}".format(query)])
print >> fwscores, "\t".join((cid, b.score))
fwscores.flush()