def bed12_to_fasta(gene_feature_file, reference_seq_file):
ref_fa = pysam.Fastafile(reference_seq_file)
for g in GeneFeature.parse(open(gene_feature_file)):
exon_seqs = []
error_occurred = False
for start, end in g.exons:
seq = ref_fa.fetch(g.chrom, start, end)
if not seq:
logging.warning(
"gene %s exon %s:%d-%d not found in reference" %
(g.tx_name, g.chrom, start, end))
error_occurred = True
break
exon_seqs.append(seq)
if error_occurred:
continue
# make fasta record
seq = ''.join(exon_seqs)
if g.strand == '-':
seq = DNA_reverse_complement(seq)
# break seq onto multiple lines
seqlines = split_seq(seq, BASES_PER_LINE)
yield (">%s range=%s:%d-%d gene=%s strand=%s\n%s" %
(GENE_REF_PREFIX + g.tx_name, g.chrom, start, end, g.gene_name,
g.strand, seqlines))
ref_fa.close()
def genepred_to_fasta(gene_feature_file, reference_seq_file):
ref_fa = pysam.Fastafile(reference_seq_file)
total = 0
used = 0
for g in GeneFeature.parse(open(gene_feature_file)):
total += 1
exon_seqs = []
error_occurred = False
for start, end in g.exons:
seq = ref_fa.fetch(g.chrom, start, end)
if (not seq) or (len(seq) < (end - start)):
logging.warning("gene %s exon %s:%d-%d not found in reference" %
(g.tx_name, g.chrom, start, end))
error_occurred = True
break
exon_seqs.append(seq)
if error_occurred:
continue
used += 1
# make fasta record
seq = ''.join(exon_seqs)
if g.strand == '-':
seq = DNA_reverse_complement(seq)
# break seq onto multiple lines
seqlines = split_seq(seq, BASES_PER_LINE)
fa_record = (">%s range=%s:%d-%d gene=%s strand=%s\n%s" %
(GENE_REF_PREFIX + g.tx_name, g.chrom, start, end,
g.gene_name, g.strand, seqlines))
yield g, fa_record
logging.info("Used %d/%d gene features" % (used,total))
ref_fa.close()
def build_gene_maps(genefile):
gene_genome_map = {}
gene_trees = collections.defaultdict(lambda: IntervalTree())
# build gene and genome data structures for fast lookup
for g in GeneFeature.parse(open(genefile)):
gene_genome_map[g.tx_name] = g
# add gene to interval tree
gene_interval = Interval(g.tx_start, g.tx_end, strand=g.strand, value=(g.tx_name))
gene_trees[g.chrom].insert_interval(gene_interval)
return gene_genome_map, gene_trees
def create_fragment_size_index(output_dir, gene_feature_file,
reference_seq_file, bowtie_build_bin,
max_fragment_size):
"""
make an alignment index containing sequences that can be used to
assess the fragment size distribution. these sequences must be
larger than the 'max_insert_size' in order to be viable for use
in characterizing the fragment size distribution.
"""
# parse genes file
genes = [g for g in GeneFeature.parse(open(gene_feature_file))]
# find all exons that are larger than the maximum estimated fragment size
exons = set([
coord for coord in find_unambiguous_exon_intervals(genes)
if (coord[2] - coord[1]) >= max_fragment_size
])
logging.info("Found %d exons larger than %d" %
(len(exons), max_fragment_size))
# extract the nucleotide sequence of the exons
logging.info("Extracting sequences to use for estimating the fragment "
" size distribution")
ref_fa = pysam.Fastafile(reference_seq_file)
frag_size_fa_file = os.path.join(output_dir, "frag_size_seq.fa")
fh = open(frag_size_fa_file, 'w')
for chrom, start, end, strand in exons:
seq = ref_fa.fetch(chrom, start, end)
if not seq:
logging.warning("exon %s:%d-%d not found in reference" %
(chrom, start, end))
continue
# make fasta record
if strand == '-':
seq = DNA_reverse_complement(seq)
# break seq onto multiple lines
seqlines = split_seq(seq, BASES_PER_LINE)
record = (">%s:%d-%d strand=%s\n%s" %
(chrom, start, end, strand, seqlines))
print >> fh, record
fh.close()
ref_fa.close()
# build bowtie alignment index from the fragment size exons
logging.info("Building bowtie index")
frag_size_index = os.path.join(output_dir, FRAG_SIZE_INDEX)
args = [bowtie_build_bin, frag_size_fa_file, frag_size_index]
return subprocess.call(args)
def build_exon_trees(samfh, genefile):
rname_tid_map = dict((rname,i) for i,rname in enumerate(samfh.references))
exon_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]
# add gene to interval tree
for start,end in g.exons[1::-1]:
exon_interval = Interval(start, end, chrom=chrom_tid, strand=g.strand, value=gene_tid)
exon_trees[chrom_tid].insert_interval(exon_interval)
return dict(exon_trees)
def create_fragment_size_index(output_dir, gene_feature_file,
reference_seq_file, bowtie_build_bin,
max_fragment_size):
"""
make an alignment index containing sequences that can be used to
assess the fragment size distribution. these sequences must be
larger than the 'max_insert_size' in order to be viable for use
in characterizing the fragment size distribution.
"""
# parse genes file
genes = [g for g in GeneFeature.parse(open(gene_feature_file))]
# find all exons that are larger than the maximum estimated fragment size
exons = set([coord for coord in find_unambiguous_exon_intervals(genes)
if (coord[2] - coord[1]) >= max_fragment_size])
logging.info("Found %d exons larger than %d" % (len(exons), max_fragment_size))
# extract the nucleotide sequence of the exons
logging.info("Extracting sequences to use for estimating the fragment "
" size distribution")
ref_fa = pysam.Fastafile(reference_seq_file)
frag_size_fa_file = os.path.join(output_dir, "frag_size_seq.fa")
fh = open(frag_size_fa_file, 'w')
for chrom, start, end, strand in exons:
seq = ref_fa.fetch(chrom, start, end)
if not seq:
logging.warning("exon %s:%d-%d not found in reference" % (chrom, start, end))
continue
# make fasta record
if strand == '-':
seq = DNA_reverse_complement(seq)
# break seq onto multiple lines
seqlines = split_seq(seq, BASES_PER_LINE)
record = (">%s:%d-%d strand=%s\n%s" %
(chrom, start, end, strand, seqlines))
print >>fh, record
fh.close()
ref_fa.close()
# build bowtie alignment index from the fragment size exons
logging.info("Building bowtie index")
frag_size_index = os.path.join(output_dir, FRAG_SIZE_INDEX)
args = [bowtie_build_bin, frag_size_fa_file, frag_size_index]
return subprocess.call(args)
def bed12_to_fasta(gene_feature_file, reference_seq_file):
ref_fa = pysam.Fastafile(reference_seq_file)
for g in GeneFeature.parse(open(gene_feature_file)):
exon_seqs = []
error_occurred = False
for start, end in g.exons:
seq = ref_fa.fetch(g.chrom, start, end)
if not seq:
error_occurred = True
break
exon_seqs.append(seq)
if error_occurred:
continue
# make fasta record
seq = ''.join(exon_seqs)
if g.strand == '-':
seq = DNA_reverse_complement(seq)
# break seq onto multiple lines
seqlines = split_seq(seq, BASES_PER_LINE)
yield (">%s range=%s:%d-%d gene=%s strand=%s\n%s" %
(config.GENE_REF_PREFIX + g.tx_name, g.chrom, start, end, g.gene_name, g.strand, seqlines))
ref_fa.close()
def build_exon_trees(samfh, genefile):
rname_tid_map = dict(
(rname, i) for i, rname in enumerate(samfh.references))
exon_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]
# add gene to interval tree
for start, end in g.exons[1::-1]:
exon_interval = Interval(start,
end,
chrom=chrom_tid,
strand=g.strand,
value=gene_tid)
exon_trees[chrom_tid].insert_interval(exon_interval)
return dict(exon_trees)