def detect_fake_chains(clones_file, Achain='A', Bchain='B'):
tcrs = parse_tsv.parse_tsv_file(
clones_file,
key_fields=[],
store_fields=['va_gene', 'cdr3a', 'vb_gene', 'cdr3b'])
fake_chains = []
if len(set([(x[0], x[1]) for x in tcrs])) == 1:
fake_chains.append(Achain)
if len(set([(x[2], x[3]) for x in tcrs])) == 1:
fake_chains.append(Bchain)
if fake_chains:
print 'Fake sequence data detected for chains: {}'.format(
' '.join(fake_chains))
return fake_chains
clones_file_with_nbrdists = '{}_nbrdists.tsv'.format(clones_file[:-4])
header = open(clones_file_with_nbrdists, 'r').readline()[:-1].split('\t')
nbrdist_tags = [
x for x in header
if x in rand_nbrdist_tags and ('wtd' in x or include_non_wtd)
]
nbrdist_tags.sort()
num_nbrdist_tags = len(nbrdist_tags)
Log('parsing {} for {} nbrdist_tags'.format(clones_file_with_nbrdists,
num_nbrdist_tags))
tcr_fields = nbrdist_tags + ['va_genes', 'vb_genes', 'cdr3a', 'cdr3b']
all_tcrs = parse_tsv.parse_tsv_file(clones_file_with_nbrdists, ['epitope'],
tcr_fields)
## look for cross-reactive tcrs
for e in all_tcrs:
new_tcrs = []
for l in all_tcrs[e]:
new_tcrs.append(
[(nbrdist_rescale * float(x)) for x in l[:num_nbrdist_tags]] + [
frozenset([
cdr3s_human.all_loopseq_representative[organism][y]
for y in l[-4].split(';')
]),
frozenset([
cdr3s_human.all_loopseq_representative[organism][y]
for y in l[-3].split(';')
]), l[-2], l[-1], False
p.flag('showmotifs')
p.flag('use_tsne')
p.multiword('epitopes').cast(lambda x: x.split())
if pngfile_prefix is None:
pngfile_prefix = clones_file[:-4]
import matplotlib
matplotlib.rcParams['mathtext.default'] = 'regular'
if not show: matplotlib.use('Agg')
import matplotlib.pyplot as plt
greek_ab = {'a': r'$\alpha$', 'b': r'$\beta$'}
all_tcrs = parse_tsv.parse_tsv_file(clones_file, ['epitope'], ['clone_id'],
True)
if showmotifs:
motifs_file = clones_file[:-4] + '_motifs.tsv'
if not exists(motifs_file):
showmotifs = False
else:
assert exists(motifs_file)
all_motifs = parse_tsv.parse_tsv_file(
motifs_file, ['epitope', 'chain'], [
'id', 'showmotif', 'chi_squared', 'matches_with_nbrs',
'matches_with_nbrs_consensus', 'expected_fraction',
'cluster_number', 'is_cluster_center', 'cluster_consensus'
])
if outfile_prefix is None:
outfile_prefix = clones_file[:-4]
import matplotlib
matplotlib.rcParams['mathtext.default'] = 'regular'
matplotlib.use('Agg')
import matplotlib.pyplot as plt
#recompute_nbrdists = True
clones_file_with_nbrdists = '{}_nbrdists.tsv'.format(clones_file[:-4])
assert exists(clones_file_with_nbrdists)
## read the epitope-specific TCRs
all_tcrs = parse_tsv.parse_tsv_file(
clones_file, ['epitope', 'subject'],
['va_genes', 'vb_genes', 'cdr3a', 'cdr3b'],
save_l=True) ## last element will be the full parse_tsv_line dictionary
if not epitopes:
epitopes = all_tcrs.keys()[:]
epitopes.sort()
Log('reading {}'.format(clones_file_with_nbrdists))
nbrdist_tag_suffix = '_wtd_nbrdist{}'.format(nbrdist_percentile)
nbrdist_tags = [
x + '_' + y + nbrdist_tag_suffix for x in epitopes for y in all_chains
]
all_nbrdists = parse_tsv.parse_tsv_file(clones_file_with_nbrdists,
['epitope', 'subject'], nbrdist_tags)
footnotes[ab+'_hydro1'] = "Mean CDR3-{} total hydrophobicity (GES scale)".format(ab)
#footnotes[ab+'_hydro2'] = "Mean total hydrophobicity (KD scale)"
footnotes[ab+'_hydro2'] = "Mean CDR3-{} total hydrophobicity (HP scale)".format(ab)
for tag in footnotes: ## no typos
assert tag in header_tags
all_dats = {}
def add_dat( epitope, tag, val ):
global all_dats
assert tag in header_tags
all_dats[ epitope ][ tag ] = val
## parse the clones file
all_tcrs = parse_tsv.parse_tsv_file( clones_file, ['epitope','subject'], ['cdr3a','cdr3b','clone_size'] )
epitopes = all_tcrs.keys()[:]
epitopes.sort()
def get_charge( cdr3 ):
return sum( ( aa_charge.get(x,0.0) for x in cdr3 ) )
def get_hp1( cdr3 ):
return sum( ( -1*GES.get(x,0.0) for x in cdr3 ) )
def get_hp2( cdr3 ):
return sum( ( HP.get(x,0.0) for x in cdr3 ) )
#return sum( ( KD.get(x,0.0) for x in cdr3 ) )
all_scores = {}