def test_junction_extraction_from_splitmap():
input = files.open(testfiles["reads_1.fastq"])
index = testfiles["genome.gem"]
gtf_junctions = set(junctions.from_gtf(testfiles["refseq.gtf"]))
jj = gem.extract_junctions(input, index, merge_with=gtf_junctions)
assert junctions is not None
assert len(jj) == 260
def test_junction_extraction_from_splitmap():
input = files.open(testfiles["reads_1.fastq"])
index = testfiles["genome.gem"]
gtf_junctions = set(junctions.from_gtf(testfiles["refseq.gtf"]))
jj = gem.extract_junctions(input, index, merge_with=gtf_junctions)
assert junctions is not None
assert len(jj) == 260
# us a mapping that preserves short indels detected during the
# extraction run.
print "Loading GTF junctions from %s" % annotation
junctions = gem.junctions.from_gtf(annotation)
# now the denovo run. This returns a tuple : (mapping, junctions)
# and here we use the merge_with parameter to merge the denovo junctions with
# the previously loaded gtf junctions.
#
# Also note that we pass only unmapped reads from the initial mapping to the junction
# extraction. This is done using the "unmapped filter"
print "Getting de-novo junctions"
(denovo_mapping, junctions) = gem.extract_junctions(
gem.filter.unmapped(initial_mapping), # only unmapped reads from the initial mapping
index,
denovo_out,
mismatches=0.04,
threads=THREADS,
merge_with=set(junctions)
)
## we filter the junnctions now by their distance and
# write all junctions with a distance <= 500000 to a file
print "Writing junctions file"
gem.junctions.write_junctions(gem.junctions.filter_by_distance(junctions, 500000), junctions_out, index)
## Initial split map run with junction sites
# Here we take all unmapped reads after the denovo junction detection and
# pass them to the split mapper along with the junctions
print "Running initial split-map"
initial_split_mapping = gem.splitmapper(