def build_tx_name_gene_map(genefile, rname_prefix=None):
rname_prefix = '' if rname_prefix is None else rname_prefix
tx_map = {}
# build gene and genome data structures for fast lookup
for g in GeneFeature.parse(open(genefile)):
tx_map[rname_prefix + g.tx_name] = g
return tx_map
def build_tid_tx_cluster_map(bamfh, line_iter, rname_prefix=None):
rname_tid_map = get_rname_tid_map(bamfh)
rname_prefix = '' if rname_prefix is None else rname_prefix
cluster_trees = collections.defaultdict(lambda: ClusterTree(0, 1))
genes = []
for g in GeneFeature.parse(line_iter):
# only use genes that are references in the sam file
rname = rname_prefix + g.tx_name
if rname not in rname_tid_map:
continue
genome_tid = rname_tid_map[g.chrom]
# insert into cluster tree
cluster_trees[genome_tid].insert(g.tx_start, g.tx_end, len(genes))
genes.append(g)
# extract gene clusters
tid_tx_cluster_map = {}
current_cluster_id = 0
for genome_tid, tree in cluster_trees.iteritems():
for start, end, indexes in tree.getregions():
# group overlapping transcripts on same strand together
strand_tx_dict = collections.defaultdict(lambda: set())
for index in indexes:
g = genes[index]
rname = rname_prefix + g.tx_name
tid = rname_tid_map[rname]
strand_tx_dict[g.strand].add(tid)
# build a map between transcript tids and all the overlapping
# transcripts on the same strand
for strand, tids in strand_tx_dict.iteritems():
for tid in tids:
tid_tx_cluster_map[tid] = current_cluster_id
current_cluster_id += 1
#print strand, [bamfh.getrname(tid) for tid in tids]
return tid_tx_cluster_map
def build_tx_name_gene_map(genefile, rname_prefix=None):
rname_prefix = '' if rname_prefix is None else rname_prefix
tx_map = {}
# build gene and genome data structures for fast lookup
for g in GeneFeature.parse(open(genefile)):
tx_map[rname_prefix + g.tx_name] = g
return tx_map
def build_gene_maps(samfh, genefile):
rname_tid_map = dict(
(rname, i) for i, rname in enumerate(samfh.references))
gene_genome_map = [None] * len(samfh.references)
gene_trees = collections.defaultdict(lambda: IntervalTree())
# build gene and genome data structures for fast lookup
for g in GeneFeature.parse(open(genefile)):
name = config.GENE_REF_PREFIX + g.tx_name
if name not in rname_tid_map:
continue
if g.chrom not in rname_tid_map:
continue
gene_tid = rname_tid_map[name]
# get reference index in sam file
chrom_tid = rname_tid_map[g.chrom]
# store gene by reference id in sam file
gene_genome_map[gene_tid] = g
# add gene to interval tree
gene_interval = Interval(g.tx_start,
g.tx_end,
chrom=g.chrom,
strand=g.strand,
value=g.tx_name)
gene_trees[chrom_tid].insert_interval(gene_interval)
return gene_genome_map, gene_trees
def build_tid_tx_cluster_map(bamfh, line_iter, rname_prefix=None):
rname_tid_map = get_rname_tid_map(bamfh)
rname_prefix = '' if rname_prefix is None else rname_prefix
cluster_trees = collections.defaultdict(lambda: ClusterTree(0,1))
genes = []
for g in GeneFeature.parse(line_iter):
# only use genes that are references in the sam file
rname = rname_prefix + g.tx_name
if rname not in rname_tid_map:
continue
genome_tid = rname_tid_map[g.chrom]
# insert into cluster tree
cluster_trees[genome_tid].insert(g.tx_start, g.tx_end, len(genes))
genes.append(g)
# extract gene clusters
tid_tx_cluster_map = {}
current_cluster_id = 0
for genome_tid, tree in cluster_trees.iteritems():
for start, end, indexes in tree.getregions():
# group overlapping transcripts on same strand together
strand_tx_dict = collections.defaultdict(lambda: set())
for index in indexes:
g = genes[index]
rname = rname_prefix + g.tx_name
tid = rname_tid_map[rname]
strand_tx_dict[g.strand].add(tid)
# build a map between transcript tids and all the overlapping
# transcripts on the same strand
for strand, tids in strand_tx_dict.iteritems():
for tid in tids:
tid_tx_cluster_map[tid] = current_cluster_id
current_cluster_id += 1
#print strand, [bamfh.getrname(tid) for tid in tids]
return tid_tx_cluster_map
def build_genome_tx_trees(genefile):
genome_tx_trees = collections.defaultdict(lambda: IntervalTree())
# build gene and genome data structures for fast lookup
for g in GeneFeature.parse(open(genefile)):
# add gene to interval tree
interval = Interval(g.tx_start, g.tx_end, strand=g.strand, value=g)
genome_tx_trees[g.chrom].insert_interval(interval)
return genome_tx_trees
def build_genome_tx_trees(genefile):
genome_tx_trees = collections.defaultdict(lambda: IntervalTree())
# build gene and genome data structures for fast lookup
for g in GeneFeature.parse(open(genefile)):
# add gene to interval tree
interval = Interval(g.tx_start, g.tx_end, strand=g.strand, value=g)
genome_tx_trees[g.chrom].insert_interval(interval)
return genome_tx_trees
def build_tid_gene_map(bamfh, genefile, rname_prefix=None):
rname_tid_map = dict((rname,tid) for tid,rname in enumerate(bamfh.references))
rname_prefix = '' if rname_prefix is None else rname_prefix
tid_tx_map = {}
# build gene and genome data structures for fast lookup
for g in GeneFeature.parse(open(genefile)):
# only use genes that are references in the sam file
rname = rname_prefix + g.tx_name
if rname not in rname_tid_map:
continue
tid = rname_tid_map[rname]
tid_tx_map[tid] = g
return tid_tx_map
def build_tid_gene_map(bamfh, genefile, rname_prefix=None):
rname_tid_map = get_rname_tid_map(bamfh)
rname_prefix = '' if rname_prefix is None else rname_prefix
tid_tx_map = {}
# build gene and genome data structures for fast lookup
for g in GeneFeature.parse(open(genefile)):
# only use genes that are references in the sam file
rname = rname_prefix + g.tx_name
if rname not in rname_tid_map:
continue
tid = rname_tid_map[rname]
tid_tx_map[tid] = g
return tid_tx_map
def build_gene_to_genome_map(line_iter, rname_prefix=None):
# create arrays to map genes in bed file to genome
rname_prefix = '' if rname_prefix is None else rname_prefix
gene_genome_map = {}
for g in GeneFeature.parse(line_iter):
rname = rname_prefix + g.tx_name
strand = 1 if g.strand == '-' else 0
exon_vectors = [(start, end) for start, end in g.exons]
if strand:
exon_vectors.reverse()
if rname in gene_genome_map:
logging.error("Duplicate references %s found in bed file" % (rname))
gene_genome_map[rname] = (g.chrom, strand, exon_vectors)
return gene_genome_map
def build_gene_to_genome_map(line_iter, rname_prefix=None):
# create arrays to map genes in bed file to genome
rname_prefix = '' if rname_prefix is None else rname_prefix
gene_genome_map = {}
for g in GeneFeature.parse(line_iter):
rname = rname_prefix + g.tx_name
strand = 1 if g.strand == '-' else 0
exon_vectors = [(start, end) for start, end in g.exons]
if strand:
exon_vectors.reverse()
if rname in gene_genome_map:
logging.error("Duplicate references %s found in bed file" %
(rname))
gene_genome_map[rname] = (g.chrom, strand, exon_vectors)
return gene_genome_map
def build_gene_interval_trees(genefile):
trees = collections.defaultdict(lambda: IntervalTree())
intervals = {}
# build gene interval trees for fast lookup by genomic position
for g in GeneFeature.parse(open(genefile)):
k = (g.chrom, g.tx_start, g.tx_end)
if k not in intervals:
# add gene to tree
txlist = []
intervals[k] = txlist
interval = Interval(g.tx_start, g.tx_end, strand=g.strand, value=txlist)
trees[g.chrom].insert_interval(interval)
else:
txlist = intervals[k]
# add isoform to value (list of isoforms that share start/end)
txlist.append(g)
return trees
def build_gene_interval_trees(genefile):
trees = collections.defaultdict(lambda: IntervalTree())
intervals = {}
# build gene interval trees for fast lookup by genomic position
for g in GeneFeature.parse(open(genefile)):
k = (g.chrom, g.tx_start, g.tx_end)
if k not in intervals:
# add gene to tree
txlist = []
intervals[k] = txlist
interval = Interval(g.tx_start, g.tx_end, strand=g.strand, value=txlist)
trees[g.chrom].insert_interval(interval)
else:
txlist = intervals[k]
# add isoform to value (list of isoforms that share start/end)
txlist.append(g)
return trees
def build_exon_interval_trees(genefile):
exon_trees = collections.defaultdict(lambda: IntervalTree())
exon_intervals = {}
# build gene and genome data structures for fast lookup
for g in GeneFeature.parse(open(genefile)):
for i, e in enumerate(g.exons):
k = (g.chrom, e[0], e[1])
if k not in exon_intervals:
# add exon to tree
txlist = []
exon_intervals[k] = txlist
interval = Interval(e[0], e[1], strand=g.strand, value=txlist)
exon_trees[g.chrom].insert_interval(interval)
else:
txlist = exon_intervals[k]
# add transcript isoform
txlist.append((g, i))
return exon_trees
def build_tid_to_genome_map(bamfh, line_iter, rname_prefix=None):
rname_tid_map = get_rname_tid_map(bamfh)
rname_prefix = '' if rname_prefix is None else rname_prefix
tid_genome_map = {}
for g in GeneFeature.parse(line_iter):
rname = rname_prefix + g.tx_name
if rname not in rname_tid_map:
continue
tid = rname_tid_map[rname]
genome_tid = rname_tid_map[g.chrom]
strand = 1 if g.strand == '-' else 0
exon_vectors = [(start, end) for start, end in g.exons]
if strand:
exon_vectors.reverse()
if tid in tid_genome_map:
logging.error("Duplicate references %s found in file" % (tid))
tid_genome_map[tid] = (genome_tid, strand, exon_vectors)
return tid_genome_map
def build_exon_interval_trees(genefile):
exon_trees = collections.defaultdict(lambda: IntervalTree())
exon_intervals = {}
# build gene and genome data structures for fast lookup
for g in GeneFeature.parse(open(genefile)):
for i,e in enumerate(g.exons):
k = (g.chrom, e[0], e[1])
if k not in exon_intervals:
# add exon to tree
txlist = []
exon_intervals[k] = txlist
interval = Interval(e[0], e[1], strand=g.strand, value=txlist)
exon_trees[g.chrom].insert_interval(interval)
else:
txlist = exon_intervals[k]
# add transcript isoform
txlist.append((g,i))
return exon_trees
def build_tid_to_genome_map(bamfh, line_iter, rname_prefix=None):
rname_tid_map = get_rname_tid_map(bamfh)
rname_prefix = '' if rname_prefix is None else rname_prefix
tid_genome_map = {}
for g in GeneFeature.parse(line_iter):
rname = rname_prefix + g.tx_name
if rname not in rname_tid_map:
continue
tid = rname_tid_map[rname]
genome_tid = rname_tid_map[g.chrom]
strand = 1 if g.strand == '-' else 0
exon_vectors = [(start, end) for start, end in g.exons]
if strand:
exon_vectors.reverse()
if tid in tid_genome_map:
logging.error("Duplicate references %s found in file" % (tid))
tid_genome_map[tid] = (genome_tid, strand, exon_vectors)
return tid_genome_map
def build_gene_maps(samfh, genefile):
rname_tid_map = dict((rname,i) for i,rname in enumerate(samfh.references))
gene_genome_map = [None] * len(samfh.references)
gene_trees = collections.defaultdict(lambda: IntervalTree())
# build gene and genome data structures for fast lookup
for g in GeneFeature.parse(open(genefile)):
name = config.GENE_REF_PREFIX + g.tx_name
if name not in rname_tid_map:
continue
if g.chrom not in rname_tid_map:
continue
gene_tid = rname_tid_map[name]
# get reference index in sam file
chrom_tid = rname_tid_map[g.chrom]
# store gene by reference id in sam file
gene_genome_map[gene_tid] = g
# add gene to interval tree
gene_interval = Interval(g.tx_start, g.tx_end, chrom=g.chrom, strand=g.strand, value=g.tx_name)
gene_trees[chrom_tid].insert_interval(gene_interval)
return gene_genome_map, gene_trees
def build_transcript_cluster_map(line_iter, rname_prefix=None):
# setup cluster trees
chrom_strand_cluster_trees = \
collections.defaultdict(lambda: {"+": ClusterTree(0,1),
"-": ClusterTree(0,1)})
transcripts = []
index_cluster_map = {}
for transcript in GeneFeature.parse(line_iter):
# insert exons into cluster tree
cluster_tree = chrom_strand_cluster_trees[transcript.chrom][transcript.strand]
i = len(transcripts)
for start,end in transcript.exons:
cluster_tree.insert(start, end, i)
# each transcript is initially in a cluster by itself
index_cluster_map[i] = set([i])
transcripts.append(transcript)
# extract gene clusters
for strand_cluster_trees in chrom_strand_cluster_trees.itervalues():
for cluster_tree in strand_cluster_trees.itervalues():
for start, end, indexes in cluster_tree.getregions():
# make new cluster by aggregating all existing
# clusters with new indexes
newclust = set(indexes)
for i in indexes:
newclust.update(index_cluster_map[i])
# map every transcript to the new cluster
for i in newclust:
index_cluster_map[i] = newclust
# enumerate all clusters
rname_prefix = '' if rname_prefix is None else rname_prefix
transcript_cluster_map = {}
for cluster_id, clust in enumerate(index_cluster_map.values()):
for i in clust:
transcript = transcripts[i]
transcript_cluster_map[rname_prefix + transcript.tx_name] = cluster_id
return transcript_cluster_map