def test_chromsizes():
assert_raises(OSError, pybedtools.get_chromsizes_from_ucsc, 'dm3', mysql='wrong path')
assert_raises(ValueError, pybedtools.get_chromsizes_from_ucsc, 'dm3', timeout=0)
try:
print pybedtools.chromsizes('dm3')
print pybedtools.get_chromsizes_from_ucsc('dm3')
assert pybedtools.chromsizes('dm3') == pybedtools.get_chromsizes_from_ucsc('dm3')
hg17 = pybedtools.chromsizes('hg17')
assert hg17['chr1'] == (0, 245522847)
fn = pybedtools.chromsizes_to_file(hg17, fn='hg17.genome')
expected = 'chr1\t245522847\n'
results = open(fn).readline()
print results
assert expected == results
# make sure the tempfile version works, too
fn = pybedtools.chromsizes_to_file(hg17, fn=None)
expected = 'chr1\t245522847\n'
results = open(fn).readline()
print results
assert expected == results
assert_raises(OSError,
pybedtools.get_chromsizes_from_ucsc,
**dict(genome='hg17', mysql='nonexistent'))
os.unlink('hg17.genome')
except OSError:
sys.stdout.write("mysql error -- test for chromsizes from UCSC didn't run")
def generate_sample_input_files(analysis, matrix):
"""Generate input files (BAM, peaks) for a sample depending on its data type."""
if analysis.data_type in REGION_BASED_DATA_TYPES:
chrom_sizes_file = tempfile.NamedTemporaryFile().name
pybedtools.get_chromsizes_from_ucsc(genome=analysis.genome,
saveas=chrom_sizes_file)
if not hasattr(analysis, "sites"):
analysis.load_data(only_these_keys=["sites"], permissive=True)
if not hasattr(analysis, "sites"):
raise AttributeError(
"Need a consensus peak set to generate sample input files.")
for sample in analysis.samples:
if hasattr(sample, "aligned_filtered_bam"):
if sample.aligned_filtered_bam is not None:
d = os.path.dirname(sample.aligned_filtered_bam)
os.makedirs(d, exist_ok=True)
generate_bam_file(
matrix.loc[:, sample.name],
sample.aligned_filtered_bam,
genome_assembly=analysis.genome,
chrom_sizes_file=chrom_sizes_file,
)
if hasattr(sample, "peaks"):
if sample.peaks is not None:
d = os.path.dirname(sample.peaks)
os.makedirs(d, exist_ok=True)
generate_peak_file(analysis.sites,
sample.peaks,
summits=False,
genome_assembly=analysis.genome)
if hasattr(sample, "summits"):
if sample.summits is not None:
d = os.path.dirname(sample.summits)
os.makedirs(d, exist_ok=True)
generate_peak_file(analysis.sites,
sample.summits,
summits=True,
genome_assembly=analysis.genome)
if hasattr(sample, "log2_read_counts"):
if sample.log2_read_counts is not None:
for res, file in sample.log2_read_counts.items():
os.makedirs(os.path.dirname(file), exist_ok=True)
generate_log2_profiles(
(2**matrix[res].loc[:, sample.name]).astype(int),
(2**matrix[res].loc[:, sample.name]
).astype(int), # this should be the background vector
file,
)
def generate_bam_file(count_vector,
output_bam,
genome_assembly="hg38",
chrom_sizes_file=None,
index=True):
"""Generate BAM file containing reads matching the counts in a vector of features"""
s = location_index_to_bed(count_vector.index)
# get reads per region
i = [i for i, c in count_vector.iteritems() for _ in range(c)]
s = s.reindex(i)
# shorten/enlarge by a random fraction; name reads
d = s["end"] - s["start"]
s = s.assign(
start=(s["start"] +
d * np.random.uniform(-0.2, 0.2, s.shape[0])).astype(int),
end=(s["end"] +
d * np.random.uniform(-0.2, 0.2, s.shape[0])).astype(int),
name=[
"{}_read_{}".format(count_vector.name, i)
for i in range(s.shape[0])
],
)
s = pybedtools.BedTool.from_dataframe(s).truncate_to_chrom(
genome=genome_assembly).sort()
# get a file with chromosome sizes (usually not needed but only for bedToBam)
if chrom_sizes_file is None:
chrom_sizes_file = tempfile.NamedTemporaryFile().name
pybedtools.get_chromsizes_from_ucsc(genome=genome_assembly,
saveas=chrom_sizes_file)
s.to_bam(g=chrom_sizes_file).saveas(output_bam)
if index:
import pysam
pysam.index(output_bam)
def test_chromsizes():
with pytest.raises(OSError):
pybedtools.get_chromsizes_from_ucsc("dm3",
mysql="wrong path",
fetchchromsizes="wrongtoo")
with pytest.raises(ValueError):
pybedtools.get_chromsizes_from_ucsc("dm3", timeout=0)
try:
print(pybedtools.chromsizes("dm3"))
print(pybedtools.get_chromsizes_from_ucsc("dm3"))
assert pybedtools.chromsizes(
"dm3") == pybedtools.get_chromsizes_from_ucsc("dm3")
hg17 = pybedtools.chromsizes("hg17")
assert hg17["chr1"] == (0, 245522847)
fn = pybedtools.chromsizes_to_file(hg17, fn="hg17.genome")
expected = "chr1\t245522847\n"
results = open(fn).readline()
print(results)
assert expected == results
# make sure the tempfile version works, too
fn = pybedtools.chromsizes_to_file(hg17, fn=None)
expected = "chr1\t245522847\n"
results = open(fn).readline()
print(results)
assert expected == results
with pytest.raises(OSError):
pybedtools.get_chromsizes_from_ucsc(**dict(
genome="hg17", mysql="nonexistent", fetchchromsizes="missing"))
os.unlink("hg17.genome")
except OSError:
sys.stdout.write(
"mysql error -- test for chromsizes from UCSC didn't run")
def enrichment(id,a, b,background, organism,name=None, score=None, strand=None, n=10, run=[]):
"""Perform enrichment analysis between two BED files.
a - path to Feature of Interest BED file (FOI)
b - path to Genomic Feature BED file (GF)
n - number of Monte-Carlo iterations
"""
write_debug("START",True)
r = {}
e = Enrichment_Par(a=a,b=b,organism=organism,n=n,background=background)
if os.path.exists(e.background):
e = e.replace(Background = BedTool(e.background))
e = e._replace(A = BedTool(str(e.a)))
e = e._replace(B = BedTool(str(e.b)))
e = e._replace(genome = pybedtools.get_chromsizes_from_ucsc(e.organism))
e = e._replace(genome_fn = pybedtools.chromsizes_to_file(e.genome))
e = e._replace(organism = str(e.organism))
flt = make_filter(name,score,strand)
e = e._replace(B = e.B.filter(flt).saveas())
e = e._replace(nA = len(e.A))
e = e._replace(nB = len(e.B))
# Exits if there are 0 GFs or 0 FOI
if not e.nA or not e.nB:
logger.info("Filter resulted in 0 Features of Interest. Terminating Run. {} (id={})".format(b,id))
return Enrichment(e.a,basename(e.b),e.nA,e.nB,"NA","NA","NA","NA","NA","NA","NA","NA","NA","NA","NA","NA","NA")
e.A.set_chromsizes(e.genome)
e.B.set_chromsizes(e.genome)
e = e._replace(obs = len(e.A.intersect(e.B, u=True)))
# This is the Monte-Carlo step. If custom background present, it is used
if 'pvalue' in run:
logger.info("Running Monte Carlo ({}): (id={})".format(b,id))
write_progress(id, "Running Monte Carlo {}".format(b))
r.update(run_montecarlo(e))
else:
r['p_value'], r['exp'] = "NA","NA"
logger.info("Skipping Monte Carlo ({}): (id={})".format(b,id))
## Uncomment to print global parameters in debug file
#write_debug("Global parameters",a = e.a,b=e.b,A= e.A,B=e.B,background = e.background,
# n=e.n,flt = e.flt,genome = e.genome,genome_fn = e.genome_fn,organism = e.organism,obs = e.obs)
# expected caluclated using pybed method CANNOT use custom background
if 'pybedtool' in run:
logger.info("Running Random Intersections ({}): (id={})".format(b,id))
write_progress(id, "Running Random Intersections: {0}".format(b))
r.update( run_pybedtool(e))
else:
r['pybedp_value'], r['pybed_exp'] = "NA","NA"
logger.info("Skipping Random Intersections")
# epected calculated using jaccard method
if 'jaccard' in run:
logger.info("Running Jaccard ({}): (id={})".format(e.b,id))
write_progress(id, "Running Jaccard {}".format(e.b))
r.update( run_jaccard(e))
else:
r['jaccardp_value'], r['jaccard_obs'],r['jaccard_exp'] = "NA","NA","NA"
logger.info("Skipping Jaccard ({}): (id={})".format(b,id))
# run kolmogorov-smornov test
if 'kolmogorov' in run:
logger.info("Running Kolmogorov-Smornov {} (id={})".format(b,id))
write_progress(id, "Running Kolmogorov-Smornov{}".format(b))
r.update( run_kolmogorov(e))
else:
r['kol_smor_p_value'] = "NA"
logger.info("Skipping Kolmogorov-Smornov {} (id={})".format(b,id))
# run proximity analysis
if 'proximity' in run:
logger.info("Running proximity {} (id={})".format(b,id))
write_progress(id, "Running proximity analysis{}".format(b))
r.update( run_proximity(e))
else:
logger.info( "Skipping Proximity")
r['obsprox'],r['expprox'],r['proximityp_value']="NA","NA", "NA"
# run hypergeometric distrubtion analysis
if 'hypergeometric' in run:
write_progress(id,"Running")
logger.info("Running hypergeometric analysis {} (id={})".format(b,id))
r.update(run_hypergeometric(e))
else:
logger.info("Skipping hypergeometric")
r['hypergeometric_p_value'] = "NA"
## Uncomment to print global parameters in debug file
#write_debug("Global parameters",a = e.a,b=e.b,A= e.A,B=e.B,background = e.background,
# n=e.n,flt = e.flt,genome = e.genome,genome_fn = e.genome_fn,organism = e.organism,obs = e.obs)
# the order of these arguments IS IMPORTANT
return Enrichment(e.a, basename(e.b), e.nA, e.nB, e.obs, r['exp'], r['p_value'],r['obsprox'],\
r['expprox'],r['pybedp_value'],r['pybed_exp'],r['jaccard_obs'],r['jaccardp_value'],\
r['jaccard_exp'],r['proximityp_value'],r['kol_smor_p_value'],r['hypergeometric_p_value'])