for l in infos:
mouse = l['subject']
epitope = l['epitope']
cdr3a = l['cdr3a']
cdr3b = l['cdr3b']
## note that we are using mm1 reps here that also dont have allele info
va_rep = l['va_label_rep']
ja_rep = l['ja_label_rep']
vb_rep = l['vb_label_rep']
jb_rep = l['jb_label_rep']
if junction_bars:
a_junction_results = tcr_sampler.analyze_junction( organism, l['va_gene'], l['ja_gene'],
cdr3a, l['cdr3a_nucseq'], return_cdr3_nucseq_src=True )
b_junction_results = tcr_sampler.analyze_junction( organism, l['vb_gene'], l['jb_gene'],
cdr3b, l['cdr3b_nucseq'], return_cdr3_nucseq_src=True )
cdr3a_new_nucseq, cdr3a_protseq_masked, cdr3a_protseq_new_nucleotide_countstring,\
a_trims, a_inserts, cdr3a_nucseq_src = a_junction_results
cdr3b_new_nucseq, cdr3b_protseq_masked, cdr3b_protseq_new_nucleotide_countstring,\
b_trims, b_inserts, cdr3b_nucseq_src = b_junction_results
## try to distinguish between N before D and N after D
for i in range(len(cdr3b_nucseq_src)):
if cdr3b_nucseq_src[i] == 'N':
if cdr3b_nucseq_src[:i].count('D')==0:
cdr3b_nucseq_src[i] = 'N1'
else:
cdr3b_nucseq_src[i] = 'N2'
else:
def make_default_logo_svg_cmds(upper_left,
width,
height,
organism,
tcr_infos,
ab,
distance_params=None,
rep_dists=None,
add_fake_alleles=False,
show_full_cdr3=False):
# right now single-chain only
# returns cmds
assert ab in 'AB'
if distance_params is None:
distance_params = tcr_distances.DistanceParams(config_string=None)
if rep_dists is None:
#print 'precomputing v-region distances'
rep_dists = tcr_distances.compute_all_v_region_distances(
organism, distance_params)
#print 'done precomputing v-region distances'
util.assign_label_reps_and_colors_based_on_most_common_genes_in_repertoire(
tcr_infos, organism)
rep_colors = {}
for info in tcr_infos:
for vj in 'vj':
for abl in 'ab':
rep = info[vj + abl + '_label_rep']
color = info[vj + abl + '_label_rep_color']
rep_colors[rep] = color
tcrs = []
dist_tcrs = []
def add_fake_allele_info(x):
if '*' not in x:
return x + '*01'
else:
return x
for l in tcr_infos:
mouse = l['subject']
epitope = l['epitope']
cdr3a = l['cdr3a']
cdr3b = l['cdr3b']
## for computing distances
va_gene = l['va_gene']
ja_gene = l['ja_gene']
va_genes = l['va_genes'].split(';')
vb_gene = l['vb_gene']
jb_gene = l['jb_gene']
vb_genes = l['vb_genes'].split(';')
# add '*01' -- hacky!
if add_fake_alleles:
va_genes = map(add_fake_allele_info, va_genes)
vb_genes = map(add_fake_allele_info, vb_genes)
va_gene = add_fake_allele_info(va_gene)
ja_gene = add_fake_allele_info(ja_gene)
vb_gene = add_fake_allele_info(vb_gene)
jb_gene = add_fake_allele_info(jb_gene)
va_reps = frozenset(
(all_genes.all_genes[organism][x].rep for x in va_genes))
vb_reps = frozenset(
(all_genes.all_genes[organism][x].rep for x in vb_genes))
dist_tcrs.append([va_reps, vb_reps, cdr3a, cdr3b])
#all_info.append( l )
## note that we are using mm1 reps here that also dont have allele info
va_rep = l['va_label_rep']
ja_rep = l['ja_label_rep']
vb_rep = l['vb_label_rep']
jb_rep = l['jb_label_rep']
cdr3a_nucseq_src = ['V'] * (3 * len(cdr3a)) ## hack, unused
cdr3b_nucseq_src = ['V'] * (3 * len(cdr3b))
if junction_bars:
if ab == 'A':
a_junction_results = tcr_sampler.analyze_junction(
organism,
va_gene,
ja_gene,
cdr3a,
l['cdr3a_nucseq'].lower(),
return_cdr3_nucseq_src=True)
cdr3a_new_nucseq, cdr3a_protseq_masked, cdr3a_protseq_new_nucleotide_countstring,\
a_trims, a_inserts, cdr3a_nucseq_src = a_junction_results
elif ab == 'B':
b_junction_results = tcr_sampler.analyze_junction(
organism,
vb_gene,
jb_gene,
cdr3b,
l['cdr3b_nucseq'].lower(),
return_cdr3_nucseq_src=True)
cdr3b_new_nucseq, cdr3b_protseq_masked, cdr3b_protseq_new_nucleotide_countstring,\
b_trims, b_inserts, cdr3b_nucseq_src = b_junction_results
## try to distinguish between N before D and N after D
for i in range(len(cdr3b_nucseq_src)):
if cdr3b_nucseq_src[i] == 'N':
if cdr3b_nucseq_src[:i].count('D') == 0:
cdr3b_nucseq_src[i] = 'N1'
else:
cdr3b_nucseq_src[i] = 'N2'
assert len(cdr3a_nucseq_src) == 3 * len(cdr3a)
assert len(cdr3b_nucseq_src) == 3 * len(cdr3b)
#print cdr3b, cdr3b_nucseq_src
tcrs.append((mouse, va_rep, ja_rep, vb_rep, jb_rep, cdr3a, cdr3b,
cdr3a_nucseq_src, cdr3b_nucseq_src, l['clone_id']))
## compute distances, used in logo construction for picking the center tcr for aligning against
#print 'computing distances:',len(dist_tcrs)
chains = ab
all_dists = np.zeros((len(dist_tcrs), len(dist_tcrs)))
for i, t1 in enumerate(dist_tcrs):
for j in range(i + 1, len(dist_tcrs)):
dist = tcr_distances.compute_distance(t1, dist_tcrs[j], chains,
rep_dists, distance_params)
all_dists[i][j] = dist
all_dists[j][i] = dist
# now make the logo
members = range(len(tcrs))
scale_w = float(width) / default_width
scale_h = float(height) / default_height
# scale everything by our desired height, width
return make_tcr_logo(upper_left, tcrs, members, all_dists, ab, rep_colors,
scale_w * default_vj_logo_width,
scale_w * default_pwmplusgaps_width,
scale_w * default_xpad, scale_h * default_pwm_height,
scale_h * default_junction_bars_height,
scale_h * default_ypad, show_full_cdr3)
if new_cdr3b_nucseq != cdr3b_nucseq: ## note note note
new_cdr3b_protseq = read_sanger_data.get_translation(
new_cdr3b_nucseq, '+1')[0]
else:
new_cdr3b_protseq = cdr3b_protseq[:]
assert new_cdr3b_protseq == read_sanger_data.get_translation(
cdr3b_nucseq, '+1')[0]
bprob_protseq = tcr_sampler.beta_cdr3_protseq_probability(
organism, vb_gene, jb_gene, new_cdr3b_protseq, verbose=verbose)
vals = dict(l) #line.split('\t') + ['']*(len(outfields)-len(infields))
if add_masked_seqs:
## junction analysis
a_junction_results = tcr_sampler.analyze_junction(
organism, va_gene, ja_gene, cdr3a_protseq, cdr3a_nucseq)
b_junction_results = tcr_sampler.analyze_junction(
organism, vb_gene, jb_gene, cdr3b_protseq, cdr3b_nucseq)
cdr3a_new_nucseq, cdr3a_protseq_masked, cdr3a_protseq_new_nucleotide_countstring,a_trims,a_inserts \
= a_junction_results
cdr3b_new_nucseq, cdr3b_protseq_masked, cdr3b_protseq_new_nucleotide_countstring,b_trims,b_inserts \
= b_junction_results
# from tcr_sampler.py:
# trims = ( v_trim, d0_trim, d1_trim, j_trim )
# inserts = ( best_d_id, n_vd_insert, n_dj_insert, n_vj_insert )
assert a_trims[1] + a_trims[2] + a_inserts[0] + a_inserts[
1] + a_inserts[2] + b_inserts[3] == 0
assert a_inserts[3] == len(cdr3a_new_nucseq)
if chain == 'A':
if correct_cdr3_seqs:
tmp_results = tcr_sampler.analyze_junction\
( organism, v_gene, j_gene, cdr3_protseq, cdr3_nucseq,
return_corrected_cdr3_seqs = True,
mismatch_score = mismatch_score_for_correcting_cdr3_seqs )
corrected_cdr3_nucseq, corrected_cdr3_protseq = list(
tmp_results)[-2:]
if corrected_cdr3_nucseq != cdr3_nucseq:
print 'fixing early sequence error', cdr3_nucseq, '==>', corrected_cdr3_nucseq
cdr3_nucseq = corrected_cdr3_nucseq[:]
cdr3_protseq = corrected_cdr3_protseq[:]
junction_results = tcr_sampler.analyze_junction(
organism, v_gene, j_gene, cdr3_protseq, cdr3_nucseq)
new_nucseq, cdr3_protseq_masked, cdr3_protseq_new_nucleotide_countstring, trims, inserts \
= junction_results
(v_trim, d0_trim, d1_trim, j_trim) = trims
(d_id, n_vd_insert, n_dj_insert, n_vj_insert) = inserts
if not new_nucseq: new_nucseq = '-'
assert d0_trim + d1_trim + n_vd_insert + n_dj_insert == 0
junction_info = '%s %s -%d -%d +%d' % (
new_nucseq, cdr3_protseq_masked, v_trim, j_trim, n_vj_insert)
else:
junction_infos = []
possible_d_ids = tcr_rearrangement_new.all_trbd_nucseq[
organism].keys()[:]