def build_subgraph(reads_in_subgraph, barcodes_unzipped): bc_file = open(barcodes_unzipped, 'rb') barcodes_iter = IO_utils.read_fastq_random( bc_file, offsets = reads_in_subgraph) subgraph_kmer_counts = Counter() while(True): try: barcode_data, _ = next(barcodes_iter) except StopIteration: break read_kmers = IO_utils.get_cyclic_kmers( barcode_data, int(args['kmer_size']), int(args['barcode_start']), int(args['barcode_end'])) for (kmer, _ ) in read_kmers: subgraph_kmer_counts[kmer] += 1 bc_file.close() edges = [] for(kmer, count) in subgraph_kmer_counts.items(): edge = Edge(kmer[0:-1], kmer[1:], count) edges.append(edge) subgraph = Graph(edges) return subgraph
def write_split_fastqs(params):
import gzip
(consensus_bcs, reads_assigned_db, reads_assigned_pipe, output_dir,
reads_unzipped, barcodes_unzipped) = params
split_dir = '%s/reads_split' % output_dir
if not os.path.exists(split_dir):
os.makedirs(split_dir)
output_files = {'batch': '%s/batch.txt' % (split_dir)}
batch_file = open(output_files['batch'], 'w')
reads_per_cell = {}
consensus_bcs.add('unassigned')
for cell in consensus_bcs:
try:
cell_offsets = IO_utils.get_from_db(reads_assigned_pipe, [cell])
except IndexError:
pass
#cell_offsets = IO_utils.read_from_pickle(reads_assigned_pickled, cell)
cell_name = 'cell_%s' % cell
#initialie all readers and writers
output_files[cell_name] = {
'reads': '%s/%s_reads.fastq.gz' % (split_dir, cell_name),
'barcodes': '%s/%s_barcodes.fastq.gz' % (split_dir, cell_name),
'umi': '%s/%s.umi.txt' % (split_dir, cell_name)
}
batch_file.write('%s\t%s\t%s\n' % \
(cell_name,
output_files[cell_name]['umi'],
output_files[cell_name]['reads']))
reads_writer = gzip.open(output_files[cell_name]['reads'], 'wb')
barcodes_writer = gzip.open(output_files[cell_name]['barcodes'], 'wb')
umi_writer = open(output_files[cell_name]['umi'], 'wb')
reads_f = open(reads_unzipped, 'rb')
barcodes_f = open(barcodes_unzipped, 'rb')
reads_iter = IO_utils.read_fastq_random(
reads_f,
offsets=[
cell_offsets[i] for i in range(len(cell_offsets)) if i % 2 == 0
])
barcodes_iter = IO_utils.read_fastq_random(
barcodes_f,
offsets=[
cell_offsets[i] for i in range(len(cell_offsets)) if i % 2 == 1
])
reads_in_cell = 0
while (True):
try:
reads_data, _ = next(reads_iter)
barcodes_data, _ = next(barcodes_iter)
reads_in_cell += 1
except StopIteration:
break
reads_data[0] += ' %s' % cell_name.replace('_', ':')
reads_data[0] = reads_data[0].replace(' ', '_')
barcodes_data[0] += ' %s' % cell_name.replace('_', ':')
barcodes_data[0] = barcodes_data[0].replace(' ', '_')
umi = barcodes_data[1][int(args['umi_start']):int(args['umi_end'])]
reads_writer.write(('\n'.join(reads_data) + '\n').encode('utf-8'))
barcodes_writer.write(
('\n'.join(barcodes_data) + '\n').encode('utf-8'))
umi_writer.write((umi + '\n').encode('utf-8'))
reads_writer.close()
umi_writer.close()
barcodes_writer.close()
reads_f.close()
barcodes_f.close()
print('\tWrote %i reads to file:\t%s' % \
(reads_in_cell, cell_name))
reads_per_cell[cell] = reads_in_cell
batch_file.close()
return output_files, reads_per_cell
