def OutputStats(self, output_fname):
fh = open(output_fname, 'w')
fh.write(
'LineageID\tLineageSizeBeforeCleaning\tNumNonTrivialSeqs\tMaxMultiplicity\tClosestV\tClosestJ\tRootId\tRootSeq\tRootCDR3\tRootDistanceFromGermline\n'
)
for l in sorted(self.clonal_lineages,
key=lambda s: len(s),
reverse=True):
vj_ann = vj_annotator.VJGeneAnnotator(l)
abundant_v = utils.GetBaseGeneName(
vj_ann.GetAbundantGene(dataset.AnnotatedGene.V))
abundant_j = utils.GetBaseGeneName(
vj_ann.GetAbundantGene(dataset.AnnotatedGene.J))
root_seq = self._GetRootSeq(l, abundant_v)
num_shms_in_root = len(
self.dataset.GetVSHMsOutsideCDR3(root_seq.id)) + len(
self.dataset.GetJSHMsOutsideCDR3(root_seq.id))
fh.write(l.id() + '\t' + str(len(l)) + '\t' +
str(self._GetNumberNonTrivialSequences(l)) + '\t' +
str(self._GetHighestMultiplicity(l)) + '\t' + abundant_v +
'\t' + abundant_j + '\t' + root_seq.id + '\t' +
root_seq.seq + '\t' +
self.dataset.GetCDR3BySeqName(root_seq.id) + '\t' +
str(num_shms_in_root) + '\n')
fh.close()
def SequenceIsGood(self, seq):
v_gene = self.full_length_lineage.Dataset().GetGeneHitBySeqName(
seq.id, dataset.AnnotatedGene.V)
j_gene = self.full_length_lineage.Dataset().GetGeneHitBySeqName(
seq.id, dataset.AnnotatedGene.J)
return utils.GetBaseGeneName(
v_gene) == self.most_freq_v and utils.GetBaseGeneName(
j_gene) == self.most_freq_j
def _GetRootSeq(self, lineage, abundant_v):
num_shms_all = sys.maxint
root_seq_all = SeqRecord(Seq(''), id='')
num_shms_prod = sys.maxint
root_seq_prod = SeqRecord(Seq(''), id='')
for seq in lineage.FullLengthSeqIdIter():
if utils.GetBaseGeneName(
self.dataset.GetGeneHitBySeqName(
seq.id, dataset.AnnotatedGene.V)) != abundant_v:
continue
cur_num_shms = len(self.dataset.GetVSHMsOutsideCDR3(seq.id)) + len(
self.dataset.GetJSHMsOutsideCDR3(seq.id))
if cur_num_shms < num_shms_all:
num_shms_all = cur_num_shms
root_seq_all = seq
cur_seq = seq.seq #self.dataset.GetCDR3BySeqName(seq.id)
aa_seq = str(Seq(cur_seq).translate())
if aa_seq.find('*') != -1:
continue
if cur_num_shms < num_shms_prod:
num_shms_prod = cur_num_shms
root_seq_prod = seq
if root_seq_prod.id != '':
return root_seq_prod
return root_seq_all
def _UpdateGeneDict(self, gene_type, lineage):
root_seq_id = lineage.RootSeqId()
gene_name = utils.GetBaseGeneName(
lineage.Dataset().GetGeneHitBySeqName(root_seq_id, gene_type))
if gene_name not in self.gene_usage[gene_type]:
self.gene_usage[gene_type][gene_name] = []
self.gene_usage[gene_type][gene_name].append(
lineage.Dataset().GetSHMsBySeqName(root_seq_id, gene_type))
def _FindMostAbundantGene(self, gene_type):
gene_dict = dict() # gene name -> num sequences
genes = []
for seq in self.full_length_lineage.FullLengthSeqIdIter():
gene_name = self.full_length_lineage.Dataset().GetGeneHitBySeqName(
seq.id, gene_type)
base_gene = utils.GetBaseGeneName(gene_name)
genes.append(base_gene)
self.gene_type_mults[gene_type] = Counter(genes)
def GetIdentifiedGenes(self):
found_d_genes = set()
for c in self.classifications:
gene_ids = c.gene_ids
if len(gene_ids) != 1:
continue
found_d_genes.add(utils.GetBaseGeneName(gene_ids[0]))
missing_d_genes = set()
for d in self.d_genes:
d_base = utils.GetBaseGeneName(d.id)
if d_base not in found_d_genes:
missing_d_genes.add(d_base)
print str(len(found_d_genes)) + " D genes are identified: " + str(
','.join([d for d in sorted(found_d_genes)]))
print str(len(self.d_genes) -
len(found_d_genes)) + ' D genes are missing: ' + str(
','.join([d for d in sorted(missing_d_genes)]))
return found_d_genes, missing_d_genes
def _FindMostAbundantGene(self, gene_type):
gene_dict = dict() # gene name -> num sequences
for seq in self.full_length_lineage.FullLengthSeqIdIter():
gene_name = self.full_length_lineage.Dataset().GetGeneHitBySeqName(
seq.id, gene_type)
base_gene = utils.GetBaseGeneName(gene_name)
if base_gene not in gene_dict:
gene_dict[base_gene] = 0
gene_dict[base_gene] += 1
self.gene_type_mults[gene_type] = gene_dict
def _FindMostAbundantGene(self, gene_type):
gene_dict = dict() # gene name -> num sequences
for seq in self.full_length_lineage.FullLengthSeqIdIter():
gene_name = self.full_length_lineage.Dataset().GetGeneHitBySeqName(seq.id, gene_type)
base_name = utils.GetBaseGeneName(gene_name)
if base_name not in gene_dict:
gene_dict[base_name] = 0
gene_dict[base_name] += 1
most_freq_gene = max(gene_dict.iteritems(), key=operator.itemgetter(1))[0]
# print "Most frequent " + str(gene_type.name) + ' gene: ' + most_freq_gene + ' (' + str(gene_dict[most_freq_gene]) + ' sequences)'
return max(gene_dict.iteritems(), key=operator.itemgetter(1))[0]
def _CreateSubstringSegmentClassification(self, segment_type, segment,
d_indices):
if segment_type == SegmentType.AMBIGUOUS_GENE:
return DSegmentClassification(segment_type, segment, '', [
utils.GetBaseGeneName(self.d_genes[ind].id)
for ind in d_indices
])
main_d_gene = self.d_genes[d_indices[0]]
segment_alignment_seq = self._GetAlignmentSeqForSubstring(
segment, main_d_gene.seq)
return DSegmentClassification(segment_type, segment_alignment_seq,
main_d_gene.seq, [main_d_gene.id])
def _GetRootSeq(self, lineage, abundant_v):
num_shms = sys.maxint
root_seq = ''
for seq in lineage.FullLengthSeqIdIter():
if utils.GetBaseGeneName(
self.dataset.GetGeneHitBySeqName(
seq.id, dataset.AnnotatedGene.V)) != abundant_v:
continue
cur_num_shms = len(self.dataset.GetVSHMsOutsideCDR3(seq.id)) + len(
self.dataset.GetJSHMsOutsideCDR3(seq.id))
if cur_num_shms < num_shms:
num_shms = cur_num_shms
root_seq = seq
return root_seq
def _SegmentIsAmbiguousGene(self, d_indices):
d_base_names = set(
[utils.GetBaseGeneName(self.d_genes[ind].id) for ind in d_indices])
return len(d_base_names) > 1
