def merge_allele_frequencies(data_folder, adaID, fragments, VERBOSE=0):
'''Merge allele frequencies at overlapping pairs'''
import warnings
import numpy as np
consensi = {frag: SeqIO.read(get_consensus_filename(data_folder, adaID, frag,
trim_primers=True), 'fasta')
for frag in fragments}
nus = {frag: np.load(get_allele_frequencies_filename(data_folder, adaID, frag))
for frag in fragments}
pairs = get_overlapping_fragments(fragments)
overlaps = {}
for (frag1, frag2) in pairs:
overlap = get_overlap(data_folder, adaID, frag1, frag2, VERBOSE=VERBOSE)
is_diff = check_overlap_consensus(data_folder, adaID, frag1, frag2,
overlap, VERBOSE=VERBOSE)
if is_diff:
warnings.warn(frag1+' and '+frag2+' have different consensi.', RuntimeWarning)
overlaps[(frag1, frag2)] = overlap
nu = []
fragments = sorted(fragments)
for i, frag in enumerate(fragments):
# If the start is not an overlap, start a new chunk and copy all
if (i == 0) or (fragments[i-1], frag) not in overlaps:
nuf = [[frag], nus[frag]]
nu.append(nuf)
# else, copy from the end of the overlap on
# FIXME: we could average the consensus zone out of indels...
else:
nuf = nu[-1]
nuf[0].append(frag)
tmp = overlaps[(fragments[i-1], frag)]
if tmp is not None:
(_, start, _) = tmp
#(recursion is not the most efficient but -- oh, well)
nuf[1] = np.concatenate([nuf[1], nus[frag][:, start:]], axis=1)
else:
tmp = np.zeros((nuf[1].shape[0], 10), float)
tmp[-1] = 1
nuf[1] = np.concatenate([nuf[1], tmp, nus[frag][:, start:]], axis=1)
return nu
def merge_consensi(data_folder, adaID, fragments, VERBOSE=0):
'''Merge consensi at overlapping pairs'''
import warnings
consensi = {frag: SeqIO.read(get_consensus_filename(data_folder, adaID, frag,
trim_primers=True), 'fasta')
for frag in fragments}
pairs = get_overlapping_fragments(fragments)
overlaps = {}
for (frag1, frag2) in pairs:
overlap = get_overlap(data_folder, adaID, frag1, frag2, VERBOSE=VERBOSE)
is_diff = check_overlap_consensus(data_folder, adaID, frag1, frag2,
overlap, VERBOSE=VERBOSE)
if is_diff:
warnings.warn(frag1+' and '+frag2+' have different consensi.', RuntimeWarning)
overlaps[(frag1, frag2)] = overlap
consensus = []
fragments = sorted(fragments)
for i, frag in enumerate(fragments):
# If the start is not an overlap, start a new consensus and copy all
if (i == 0) or (fragments[i-1], frag) not in overlaps:
cons = [[frag], str(consensi[frag].seq)]
consensus.append(cons)
# copy from the end of the overlap on
else:
cons = consensus[-1]
cons[0].append(frag)
tmp = overlaps[(fragments[i-1], frag)]
if tmp is not None:
(_, start, _) = tmp
cons[1] = cons[1]+str(consensi[frag][start:].seq)
else:
cons[1] = cons[1]+('N' * 10)+str(consensi[frag].seq)
# Make SeqRecords out of consensi
for i, (frags, cons) in enumerate(consensus):
name = 'adaID_'+str(adaID)+'_'+'-'.join(frags)
rec = SeqRecord(Seq(cons, IUPAC.ambiguous_dna),
id=name, name=name)
consensus[i] = (frags, rec)
return consensus
