def addplot(oindexlist, ofracslist, n_seqs, fname, title):
hist = Hist(30, 0., 1.)
for ofracs in ofracslist:
hist.fill(ofracs)
fig, ax = self.plotting.mpl_init()
hist.mpl_plot(ax, remove_empty_bins=True)
ax.text(0.65,
0.8 * ax.get_ylim()[1],
'size: %d' % n_seqs,
fontsize=20,
fontweight='bold')
ax.text(0.65,
0.7 * ax.get_ylim()[1],
'h: %.2f' %
utils.fay_wu_h(line=None,
restrict_to_region=restrict_to_region,
occurence_indices=oindexlist,
n_seqs=n_seqs),
fontsize=20,
fontweight='bold')
regionstr = restrict_to_region + ' ' if restrict_to_region is not None else ''
self.plotting.mpl_finish(
ax,
plotdir,
fname,
title=title,
xlabel=regionstr + 'mutation frequency',
ylabel=regionstr + 'density of mutations',
xticks=[0, 1],
log=''
) # xticks=[min(occurence_fractions), max(occurence_fractions)],
self.addfname(fnames, fname)
def print_stuff(line):
cluster_index = sorted_clusters.index(cluster)
naive_cdr3, matureiseq0_cdr3 = utils.subset_sequences(
line, iseq=0, restrict_to_region='cdr3'
) # line['naive_seq'][(line['codon_positions']['v']):((line['codon_positions']['j'])+3)] #get nt sequence of CDR3 from first base of cysteine through last base of tryptophan
# mature_cdr3_seqs = [] # trying to translate the consensus cdr3 so I can search these with my seed seqs
# for iseq in range(len(line['unique_ids'])):
# naive_cdr3_seq, mature_cdr3_seq = utils.subset_sequences(line, iseq=iseq, restrict_to_region='cdr3')
# mature_cdr3_seqs.append(mature_cdr3_seq)
# translated_cdr3 = Seq().... not done
cdr3_aa = '%-30s' % Seq(naive_cdr3).translate()
if any('-ig' in s for s in line['unique_ids']):
cdr3_aa = utils.color('red', cdr3_aa, width=30)
print '%4s %s %s %s %5d %5d %5d %7.3f %8.4f %2d %s %4.2f' % (
cluster_index,
utils.color_gene(line['v_gene'], width=15),
utils.color_gene(line['d_gene'], width=15),
utils.color_gene(line['j_gene'], width=10),
len(line['unique_ids']),
numpy.mean(line['n_mutations']),
numpy.median(line['n_mutations']),
numpy.mean(line['mut_freqs']),
float(len(cluster)) / n_total,
(line['cdr3_length'] / 3),
cdr3_aa,
utils.fay_wu_h(line, debug=False),
)
def print_stuff(line):
cluster_index = sorted_clusters.index(cluster)
naive_cdr3, matureiseq0_cdr3 = utils.subset_sequences(line, iseq=0, restrict_to_region='cdr3') # returns the CDR3 nt sequence for naive, and the first mutated sequence (iseq0); CDR3 = first base of cysteine through last base of tryptophan
# mature_cdr3_seqs = [] # trying to translate the consensus cdr3 so I can search these with my seed seqs
# for iseq in range(len(line['unique_ids'])):
# naive_cdr3_seq, mature_cdr3_seq = utils.subset_sequences(line, iseq=iseq, restrict_to_region='cdr3')
# mature_cdr3_seqs.append(mature_cdr3_seq)
# mature_cdr3_seqs
# translated_cdr3 = mature_cdr3_seqs.translate()
cdr3_aa = '%-30s' % Seq(naive_cdr3).translate()
# If a cluster contains one of our seed seqs, color this CDR3 red
if any('-ig' in s for s in line['unique_ids']):
cdr3_aa = utils.color('red', cdr3_aa, width=30)
if args.cdr3 in cdr3_aa: # Only print clusters with naive CDR3 that matches our specified --cdr3 argument
print 'index genes size n muts SHM rep frac CDR3 FayWuH'
print ' mean med len seq'
print '%4s %s %s %s %5d %5d %5d %7.3f %8.4f %2d %s %4.2f' % (
cluster_index,
utils.color_gene(line['v_gene'], width=15),
utils.color_gene(line['d_gene'], width=15),
utils.color_gene(line['j_gene'], width=10),
len(line['unique_ids']),
numpy.mean(line['n_mutations']),
numpy.median(line['n_mutations']),
numpy.mean(line['mut_freqs']),
float(len(cluster)) / n_total,
(line['cdr3_length']/3),
cdr3_aa,
utils.fay_wu_h(line, debug=False),
)
# print 'number of mutations per sequence in cluster', sorted(line['n_mutations'])
print len(line['naive_seq']), 'length of naive seq'
# utils.print_reco_event(utils.synthesize_single_seq_line(line, iseq=0)) # print ascii-art representation of the rearrangement event
print 'unique_ids: ', getkey(line['unique_ids'])
print
print utils.print_reco_event(line)
def print_stuff(line):
intscore = 0 # create a clonal family scoring system
cluster_index = sorted_clusters.index(cluster)
shm_index = shm_clusters.index(cluster)
naive_cdr3, matureiseq0_cdr3 = utils.subset_sequences(
line, iseq=0, restrict_to_region='cdr3'
) # line['naive_seq'][(line['codon_positions']['v']):((line['codon_positions']['j'])+3)] #get nt sequence of CDR3 from first base of cysteine through last base of tryptophan
# mature_cdr3_seqs = [] # trying to translate the consensus cdr3 so I can search these with my seed seqs
# for iseq in range(len(line['unique_ids'])):
# naive_cdr3_seq, mature_cdr3_seq = utils.subset_sequences(line, iseq=iseq, restrict_to_region='cdr3')
# mature_cdr3_seqs.append(mature_cdr3_seq)
# translated_cdr3 = Seq().... not done
cdr3_aa = '%-30s' % Seq(naive_cdr3).translate()
if any('-ig' in s for s in line['unique_ids']):
cdr3_aa = utils.color('red', cdr3_aa, width=30)
# score clusters based on cluster size
if cluster_index < 25:
intscore = intscore + 4
elif cluster_index >= 25 and cluster_index <= 50:
intscore = intscore + 3
elif cluster_index >= 50 and cluster_index <= 75:
intscore = intscore + 2
elif cluster_index >= 75 and cluster_index <= 100:
intscore = intscore + 1
# score clusters based on SHM
if shm_index < 25:
intscore = intscore + 4
elif shm_index >= 25 and shm_index <= 50:
intscore = intscore + 3
elif shm_index >= 50 and shm_index <= 75:
intscore = intscore + 2
elif shm_index >= 75 and shm_index <= 100:
intscore = intscore + 1
# score clusters based on SFS
if utils.fay_wu_h(line, debug=False) <= -20:
intscore = intscore + 4
elif utils.fay_wu_h(line, debug=False) <= -10:
intscore = intscore + 3
elif utils.fay_wu_h(line, debug=False) <= 0:
intscore = intscore + 2
elif utils.fay_wu_h(line, debug=False) <= 10:
intscore = intscore + 1
# score by bnAb gene usage
if (line['v_gene']).split('*')[0] in (
cd4bs_genes or glycan_genes or bridging_genes or mper_genes
): # beware this does not include CDR3 length of bnAb VH genes
intscore = intscore + 4
print '%4s %4s %s %s %s %5d %5d %5d %7.3f %8.4f %2d %s %4.2f' % (
intscore,
cluster_index,
utils.color_gene(line['v_gene'], width=15),
utils.color_gene(line['d_gene'], width=15),
utils.color_gene(line['j_gene'], width=10),
len(line['unique_ids']),
numpy.mean(line['n_mutations']),
numpy.median(line['n_mutations']),
numpy.mean(line['mut_freqs']),
float(len(cluster)) / n_total,
(line['cdr3_length'] / 3),
cdr3_aa,
utils.fay_wu_h(line, debug=False),
)
sorted_clusters = sorted(annotations,
key=lambda q: len(annotations[q]['unique_ids']),
reverse=True)
#### sorted_clusters = [c for c in sorted_clusters if utils.is_functional(annotations[c])] # checks if the cluster contains ANY non-functional sequences
# total size of repertoire (number sequences)
n_total = sum([len(cluster) for cluster in sorted_clusters])
# add more criteria
biggest_clusters = sorted_clusters[:100] # 100 biggest clusters
shm_clusters = sorted(biggest_clusters,
key=lambda q: numpy.mean(annotations[q]['mut_freqs']),
reverse=True) # rank by SHM
sfs_clusters = sorted(biggest_clusters,
key=lambda q: utils.fay_wu_h(annotations[q], debug=False)
) # rank by SFS score
cluster_sfses = {}
for cluster in biggest_clusters:
cluster_sfses[cluster] = utils.fay_wu_h(annotations[cluster], debug=False)
print numpy.mean(cluster_sfses.values())
print numpy.std(cluster_sfses.values())
print numpy.percentile(cluster_sfses.values(), 5)
print numpy.percentile(cluster_sfses.values(), 10)
print numpy.percentile(cluster_sfses.values(), 50)
print numpy.percentile(cluster_sfses.values(), 80)
print numpy.percentile(cluster_sfses.values(), 90)
# create function that gives me the score - this function calls a subfunction for each metric (i.e. percentile). The superfunction can then weight the metrics
# give 0 points to anyone not in top 30 percentile
# sort by size
sorted_clusters = sorted(annotations,
key=lambda q: len(annotations[q]['unique_ids']),
reverse=True)
# sorted_clusters = [c for c in sorted_clusters if utils.is_functional(annotations[c])] # checks if the cluster contains ANY non-functional sequences
n_total = sum([len(cluster) for cluster in sorted_clusters])
# add more criteria
biggest_clusters = sorted_clusters[:100] # 100 biggest clusters
shm_clusters = sorted(biggest_clusters,
key=lambda q: numpy.mean(annotations[q]['mut_freqs']),
reverse=True)
sfs_clusters = sorted(
biggest_clusters,
key=lambda q: utils.fay_wu_h(annotations[q], debug=False))
# cluster size: print x biggest clusters
print '\x1b[1;32;40m' + ' printing the largest clusters' + '\x1b[0m'
for cluster in sorted_clusters[:5]:
# if sorted_clusters.index(cluster) < 50:
# print_stuff(annotations[cluster])
print_stuff(annotations[cluster])
# high mean %SHM: print most mutated clusters from 100 biggest clusters
mutclust = int(args.nclust)
print '\x1b[1;32;40m' + ' printing the most mutated clusters (within 100 biggest)' + '\x1b[0m'
for cluster in shm_clusters[:mutclust]:
# if sorted_clusters.index(cluster) < 50:
# print_stuff(annotations[cluster])
print_stuff(annotations[cluster])