win_start = start
while win_start + width - gap < min(L, end):
win_end = min(win_start + width, end, L)
if VERBOSE >= 1:
print patient.code, win_start, win_end
if VERBOSE >= 2:
print 'Get region haplotypes'
try:
datum = patient.get_local_haplotype_count_trajectories(\
'genomewide',
start=win_start,
end=win_end,
filters=['noN',
'mincount='+str(countmin),
'freqmin='+str(freqmin),
],
VERBOSE=VERBOSE,
align=True,
return_dict=True)
except RoiError:
win_start += gap
continue
if not len(datum['ind']):
win_start += gap
continue
datum['times'] = patient.times[datum['ind']]
datum['pcode'] = patient.code
for pname, patient in patients.iterrows():
patient = Patient(patient)
patient.discard_nonsequenced_samples()
if VERBOSE >= 1:
print patient.name, roi
if VERBOSE >= 2:
print 'Get haplotype trajectories'
try:
(ht, indt, htseqs) = patient.get_region_count_trajectories(roi[0],
VERBOSE=VERBOSE)
except IOError:
(ht, indt, htseqs) = patient.get_local_haplotype_count_trajectories(roi,
VERBOSE=VERBOSE)
if VERBOSE >= 2:
print 'Align haplotypes and delete rare ones'
indht = (ht > 5).any(axis=0)
ht = ht[:, indht]
htseqs = htseqs[indht]
# Eliminate time points that are empty after this filter
ind_keep = ht.any(axis=1)
indt = indt[ind_keep]
ht = ht[ind_keep]
# Prepare data structures, both the matrix with sequences and the hft
hft = (1.0 * ht.T / ht.sum(axis=1)).T
alim = np.array(build_msa(htseqs))
win_start = start
while win_start + width - gap < min(L, end):
win_end = min(win_start + width, end, L)
if VERBOSE >= 1:
print patient.code, win_start, win_end
if VERBOSE >= 2:
print 'Get region haplotypes'
try:
datum = patient.get_local_haplotype_count_trajectories(\
'genomewide',
start=win_start,
end=win_end,
filters=['noN',
'mincount='+str(countmin),
'freqmin='+str(freqmin),
],
VERBOSE=VERBOSE,
align=True,
return_dict=True)
except RoiError:
win_start += gap
continue
if not len(datum['ind']):
win_start += gap
continue
datum['times'] = patient.times[datum['ind']]
datum['pcode'] = patient.code
patient = Patient(patient)
patient.discard_nonsequenced_samples()
if VERBOSE >= 1:
print patient.name, roi
if VERBOSE >= 2:
print 'Get haplotype trajectories'
try:
(ht, indt,
htseqs) = patient.get_region_count_trajectories(roi[0],
VERBOSE=VERBOSE)
except IOError:
(ht, indt,
htseqs) = patient.get_local_haplotype_count_trajectories(
roi, VERBOSE=VERBOSE)
if VERBOSE >= 2:
print 'Align haplotypes and delete rare ones'
indht = (ht > 5).any(axis=0)
ht = ht[:, indht]
htseqs = htseqs[indht]
# Eliminate time points that are empty after this filter
ind_keep = ht.any(axis=1)
indt = indt[ind_keep]
ht = ht[ind_keep]
# Prepare data structures, both the matrix with sequences and the hft
hft = (1.0 * ht.T / ht.sum(axis=1)).T
alim = np.array(build_msa(htseqs))
