def load_transcript_strands(loc):
transcript_strand = {}
filename = loc + 'transcript.txt.gz'
with fast_gzip_read(filename) as f:
for line in tqdm(f, total=count_lines(filename)):
data = line.split('\t')
transcript_strand[data[14]] = int(data[6])
return transcript_strand
def load_chromosome_and_region_names(loc):
seq_region = {}
# seq_region_id, name, cord_system_fk
filename = loc + 'seq_region.txt.gz'
with fast_gzip_read(filename) as f:
for line in tqdm(f, total=count_lines(filename)):
data = line.split('\t')
seq_region[int(data[0])] = data[1]
return seq_region
def load_variation_sources(loc):
sources = {}
# 'source_id', 'name', 'version', 'description', 'url', 'type', 'somatic_status', 'data_types'
filename = loc + 'source.txt.gz'
with fast_gzip_read(filename) as f:
for line in tqdm(f, total=count_lines(filename)):
data = line.split('\t')
sources[int(data[0])] = data[1]
gc.collect()
return sources
def get_cds_positions(transcripts):
cds_positions = {}
with fast_gzip_read('ucsc/ref_gene.tsv.gz') as f:
header = next(f)
# assert header == '#bin name chrom strand txStart txEnd cdsStart cdsEnd exonCount exonStarts exonEnds score name2 cdsStartStat cdsEndStat exonFrames'
for line in f:
data = line.split('\t')
refseq = data[1]
if refseq not in transcripts:
continue
start, end = map(int, data[6:7 + 1])
cds_positions[refseq] = (start, end)
return cds_positions
def __init__(self, filename=None):
data = defaultdict(list)
if not filename:
filename = self.filename
if not filename:
raise ValueError
with fast_gzip_read(filename, processes=6) as f:
header = next(f)
assert header == '#hg19.knownToRefSeq.value hg19.knownToEnsembl.value\n'
for line in f:
line = line.strip().split('\t')
try:
ref_id, unknown_id = line
except ValueError:
continue
if unknown_id != 'n/a':
data[ref_id].append(unknown_id)
self.data = data
def _get_all_zscores():
zscores = []
from multiprocess import fast_gzip_read
print('Counting...')
count = count_spidex()
print('Loading...')
with fast_gzip_read(SPIDEX_LOCATION) as f:
header = next(f)
get_dpsi_zscore = itemgetter(headers.index('dpsi_zscore'))
for line in tqdm(f, total=count - 1):
try:
data = line.rstrip('\n').split('\t')
# record = SpidexRecord(*data)
# zscores.append(record.dpsi_zscore)
zscores.append(float(get_dpsi_zscore(data)))
except Exception as e:
print(e)
continue
return zscores
def import_expressed_genes(bdb,
tissues=GTEX_TISSUES,
path=DEFAULT_PATH,
suffix=DEFAULT_GENE_SUFFIX):
print('Importing expressed genes:')
count = count_all(tissues, path, suffix)
with tqdm(total=count) as progress:
for tissue_name in tissues:
file_name = tissue_name + suffix
file_path = os.path.join(path, file_name)
print('Loading', file_name)
with fast_gzip_read(file_path) as file_object:
# skip header
next(file_object)
for line in file_object:
data = line.split()
"""
gene_id: (
gene_name,
gene_chr,
gene_start,
gene_end,
strand,
)
"""
if not bdb[data[0]]:
bdb[data[0]].extend(data[1:6])
else:
assert bdb[data[0]] == data[1:6]
progress.update(1)
def iterate_over_expression(tissues_list=GTEX_TISSUES,
path=DEFAULT_PATH,
suffix=DEFAULT_SUFFIX):
for tissue_name in tissues_list:
file_name = tissue_name + suffix
file_path = os.path.join(path, file_name)
print('Loading', file_name)
with fast_gzip_read(file_path, processes='all') as file_object:
header_line = next(file_object)
header = dict()
for position, name in enumerate(header_line.split()):
header[name] = position
slope_pos = header['slope']
gene_id_pos = header['gene_id']
variant_id_pos = header['variant_id']
for line in file_object:
data = line.split()
yield (data[variant_id_pos], tissue_name, data[slope_pos],
data[gene_id_pos])