def calc_permiscuity(c, juncmap5p, juncmap3p, ggmap):
# subtract one since 5' junc position is an open interval
coord5p = gene_to_genome_pos(c.mate5p.tx_name, c.mate5p.end - 1, ggmap)
coord3p = gene_to_genome_pos(c.mate3p.tx_name, c.mate3p.start, ggmap)
partners = juncmap5p[coord5p]
cov = partners[coord3p]
total_cov = sum(partners.itervalues())
frac5p = cov / float(total_cov)
partners = juncmap3p[coord3p]
cov = partners[coord5p]
total_cov = sum(partners.itervalues())
frac3p = cov / float(total_cov)
return frac5p, frac3p
def build_junc_permiscuity_map(chimeras, ggmap):
junc5p_map = collections.defaultdict(lambda: collections.defaultdict(lambda: 0))
junc3p_map = collections.defaultdict(lambda: collections.defaultdict(lambda: 0))
for c in chimeras:
# subtract one since 5' junc position is an open interval
coord5p = gene_to_genome_pos(c.mate5p.tx_name, c.mate5p.end - 1, ggmap)
coord3p = gene_to_genome_pos(c.mate3p.tx_name, c.mate3p.start, ggmap)
# keep track of total reads eminating from each 5' junction
# by keeping a dictionary for each 5' junction to all 3' junctions
# that stores the maximum coverage at that 5'/3' pair
partners = junc5p_map[coord5p]
count = partners[coord3p]
partners[coord3p] = max(count, c.weighted_cov)
# repeat for 3' partner
partners = junc3p_map[coord3p]
count = partners[coord5p]
partners[coord5p] = max(count, c.weighted_cov)
#print '5P', c.mate5p.gene_name, len(partners), sum(partners.itervalues())
#print '3P', c.mate3p.gene_name, len(partners), sum(partners.itervalues())
return junc5p_map, junc3p_map
def build_junc_coverage_map(chimeras, ggmap):
junc_cov_map = collections.defaultdict(lambda: [None, None, None])
num_chimeras = 0
for c in chimeras:
num_chimeras += 1
# convert to genomic coords
# subtract one since 5' junc position is an open interval
coord5p = gene_to_genome_pos(c.mate5p.tx_name, c.mate5p.end - 1, ggmap)
coord3p = gene_to_genome_pos(c.mate3p.tx_name, c.mate3p.start, ggmap)
# keep track of maximum coverage isoform
pairkey = (coord5p, coord3p)
paircov = (c.encomp_and_spanning, c.weighted_cov, c.encomp_or_spanning)
data = junc_cov_map[pairkey]
if (data[0] is None) or (cmp(paircov, data[0]) > 0):
# store encomp/spanning, weighted coverage, and total reads
data[0] = paircov
data[1] = c.mate5p.tx_name
data[2] = c.mate3p.tx_name
logging.debug("Parsed %d chimeras" % (num_chimeras))
kept_isoforms = set(tuple(v[1:3]) for v in junc_cov_map.itervalues())
#del junc_cov_map
logging.debug("Kept %d highest coverage isoforms" % (len(kept_isoforms)))
return kept_isoforms
def build_junc_coverage_map(chimeras, ggmap):
junc_cov_map = collections.defaultdict(lambda: [None, None, None])
num_chimeras = 0
for c in chimeras:
num_chimeras += 1
# convert to genomic coords
# subtract one since 5' junc position is an open interval
coord5p = gene_to_genome_pos(c.mate5p.tx_name, c.mate5p.end - 1, ggmap)
coord3p = gene_to_genome_pos(c.mate3p.tx_name, c.mate3p.start, ggmap)
# keep track of maximum coverage isoform
pairkey = (coord5p, coord3p)
paircov = (c.encomp_and_spanning, c.weighted_cov, c.encomp_or_spanning)
data = junc_cov_map[pairkey]
if (data[0] is None) or (cmp(paircov, data[0]) > 0):
# store encomp/spanning, weighted coverage, and total reads
data[0] = paircov
data[1] = c.mate5p.tx_name
data[2] = c.mate3p.tx_name
logging.debug("Parsed %d chimeras" % (num_chimeras))
kept_isoforms = set(tuple(v[1:3]) for v in junc_cov_map.itervalues())
#del junc_cov_map
logging.debug("Kept %d highest coverage isoforms" % (len(kept_isoforms)))
return kept_isoforms
