def __missing__(self, k):
self[k] = [
seq_to_numpy(s)
for s
in convert_seqs(sequence_files[k])
]
return self[k]
def __missing__(self, key):
dataset, cross_fold_idx = key
positive_seqs = [
seq_to_numpy(s)
for s
in convert_seqs(sequence_files['%s-%d-validate' % (dataset, cross_fold_idx)])
]
# load the negative set and match the lengths of the positive sequences
negative_seqs = [
seq_to_numpy(neg[:len(pos)])
for pos, neg
in zip(
positive_seqs,
self.negative_seq_generators[dataset]
)
]
if len(negative_seqs) != len(positive_seqs):
raise RuntimeError('Not enough sequences in negative set to match positive sequences. %d positive, %d negative sequences' % (len(positive_seqs), len(negative_seqs)))
for i, (pos, neg) in enumerate(zip(positive_seqs, negative_seqs)):
if len(neg) < len(pos):
raise RuntimeError(
'Not enough bases in negative sequence %d to match length of positive sequence: positive sequence has %d bases and negative sequence has %d bases' % (
i,
pos.shape[0],
neg.shape[0]
)
)
self[key] = (positive_seqs, negative_seqs)
return self[key]
K_min,
K_max,
max_mismatches,
)
if os.system(cmd):
raise RuntimeError("Could not run risotto with args: %s" % cmd)
return _parse_output(output_file)
finally:
if os.access(alphabet_file, os.R_OK):
os.remove(alphabet_file)
if os.access(fasta_file, os.R_OK):
os.remove(fasta_file)
if os.access(output_file, os.R_OK):
os.remove(output_file)
if "__main__" == __name__:
logging.basicConfig(level=logging.DEBUG)
from sequences import convert_seqs, rev_comp, numpy_to_seq, seq_to_numpy
test_seq_file = "synthetic-2/synthetic-sequences-K10-g0.50-N200-L200-seed4-0.fa"
sequences = convert_seqs(test_seq_file)
# sequences.extend([numpy_to_seq(rev_comp(seq_to_numpy(s))) for s in sequences])
# sequences[0] = sequences[0].replace('a','n')
results = risotto(sequences, max_mismatches=1)
results.sort(key=lambda x: x[2]) # sort by number of sequences that motif is found in
print results[-10:]