class CompoundHets(object):
def __init__(self, input, ped, frequency_cutoff=None, consequence=None, output=None, prefix=None, transcript=False, debug=False,
gene_lookback=5, force=False):
self.frequency_cutoff = frequency_cutoff
self.consequence = consequence
self.transcript = transcript
self.output = output
self.prefix = prefix
self.force = force
self.aggregate_file = os.path.join(output, 'compound_hets.aggregate.tsv')
if self.prefix:
self.aggregate_file = os.path.join(output, self.prefix + '.compound_hets.aggregate.tsv')
self.debug = debug
self.Utils = Utilities(self.frequency_cutoff, self.consequence, self.transcript, self.debug)
self.families = self.Utils.read_ped_file(ped)
self.lookback = gene_lookback
self.setup_output(output, self.families)
self.process_file(input)
# Print each compound het pair for each sample to a sample file
# Input:
# samples: dictionary of lists of tuples of Variants
# output: output prefix
def print_samples(self, samples, output):
if not samples:
return
for s in samples:
filename = "%s.compound_het_pairs.txt" % s
if self.prefix:
filename = self.prefix + "." + filename
if output:
filename = "%s.%s.compound_het_pairs.txt" % (output, s)
f = open(filename, 'w')
f.write("#PAIR\tCHR\tPOS\tGENE\n")
count = 1
for pair in samples[s]:
v1 = pair[0]
v2 = pair[1]
f.write(v1.for_sample_file(count) + "\n")
f.write(v2.for_sample_file(count) + "\n")
count += 1
f.close()
# Print all pairs of compound hets to file
# Input:
# positions: A list of tuples of variants
# output: output file prefix
def print_positions(self, positions, output):
if not positions:
return
filename = "all_compound_het_pairs.txt"
if output:
filename = "%s.all_compound_het_pairs.txt" % output
f = open(filename, 'w')
f.write("#PAIR\tCHR\tPOS\tGENE\tFROM_MOTHER\tFROM_FATHER\n")
count = 1
for p in positions:
v1 = p[0]
v2 = p[1]
f.write(v1.for_variant_file(count) + "\n")
f.write(v2.for_variant_file(count) + "\n")
count += 1
f.close()
# Parse each line in the input file, convert it to a Variant object, and assign it to a gene within the gene
# dictionary
# Input:
# input: input filename
# Return:
# genes: dictionary of lists of Variants
def process_file(self, input):
current_genes = defaultdict(list)
gene_list = []
checked_genes = []
with open(input) as f:
for line in f:
if line.startswith('#'):
continue
variant = self.Utils.parse_line(line)
if variant is None:
continue
if variant.gene not in current_genes:
if variant.gene in checked_genes:
print("%s ALREADY CHECKED! File may be out of order!" % variant.gene)
else:
checked_genes.append(variant.gene)
gene_list.append(variant.gene)
# Keep a lookback of 5 genes to account for gene overlaps
if len(current_genes) > self.lookback:
to_process = gene_list.pop(0)
self.process_gene(current_genes[to_process])
del current_genes[to_process]
current_genes[variant.gene].append(variant)
for gene in current_genes:
self.process_gene(current_genes[gene])
# Get any het calls for a gene and print the results
# Input:
# gene: A list of variants (Variant objects) within a gene
# Return:
# Nothing
def process_gene(self, gene):
if self.debug:
print("Processing gene")
hets = self.check_gene(gene)
if hets:
self.print_hets(hets)
# For each gene in the gene list, check whether each sample has a compound het pair within the gene
# Input:
# genes: dictionary of lists of Variants
# Return:
# samples: dictionary of lists of tuples of Variants
# positions: A list of tuples of variants
def check_gene(self, gene):
hets = []
identified_pairs = []
for v1 in gene:
for patient_id in v1.from_father_affected + v1.from_father_unaffected:
family_id = self.Utils.patient_id_to_family(patient_id)
if not self.families[family_id].members[patient_id].has_disease():
continue
for v2 in gene:
if v1.pos == v2.pos:
continue
if v2.has_maternal(patient_id):
key = "-".join(sorted([v1.pos, v2.pos, self.Utils.patient_id_to_family(patient_id)]))
het = CompoundHet(patient_id, v2, v1, self.families)
if not key in identified_pairs:
if self.transcript:
if v1.transcript == v2.transcript:
hets.append(het)
identified_pairs.append(key)
else:
hets.append(het)
identified_pairs.append(key)
return hets
# Print out the hets - debugging function
# input:
# hets: a CompoundHet object
def print_hets(self, hets):
for h in hets:
h.print_aggregate(self.aggregate_file)
h.print_family(self.output, self.prefix)
# Create output folder and initialize all output files with headers
# input:
# output: the output directory
# families: A dictionary of Family objects. Family IDs are the keys of the dictionary.
def setup_output(self, output, families):
# Due to popular demand this has been removed.
#if os.path.exists(output) and not self.force:
# print("Output folder found! Please choose a different output path.")
# exit(0)
#elif os.path.exists(output) and self.force:
# print("Existing output folder being moved to %s" % output + ".old")
# if os.path.exists(output + ".old"):
# print("Removing existing old output: %s" % output + ".old")
# #os.remove(output + ".old")
# shutil.rmtree(output + ".old")
# os.rename(output, output + ".old")
if not os.path.exists(output):
os.mkdir(output)
file = open(self.aggregate_file,'w')
print('Gene Maternal_VarID Maternal_Var_CSQ Paternal_VarID Paternal_Var_CSQ '
'Maternal_Var_CSQ-Paternal_Var_CSQ family fam_n_children fam_n_aff fam_n_unaff fam_n_missing_aff'
' fam_n_missing_unaff n_uncertain_aff n_uncertain_unaff n_aff_carriers n_unaff_carriers '
'n_noncarrier_aff n_noncarrier_unaff frac_of_aff frac_of_unaff frac_of_aff_missing_adjusted '
'frac_of_unaff_missing_adjusted frac_of_aff_uncertain_adjusted frac_of_unaff_uncertain_adjusted '
'frac_of_aff_carriers_missing_uncertain_adj frac_of_unaff_carriers_missing_uncertain_adj '
'Info_Variant1 Info_Variant2', file=file)
file.close()
for f in families:
filename = f + '.family_file.tsv'
if self.prefix:
filename = self.prefix + "." + filename
family_file = os.path.join(output, filename)
file = open(family_file,'w')
print('Gene Maternal_Var_CSQ-Paternal_Var_CSQ family child_id is_aff Inheritance_Variant1 Chr '
'Position Ref Alt VariantID esp6500siv2_all ExAC_ALL ThousandGenomes_2014oct_all cg46 CADD'
' CADD_Phred Polyphen2_HDIV_score Polyphen2_HDIV_pred Polyphen2_HVAR_score Polyphen2_HVAR_pred'
' consequence Inheritance_Variant2 Chr Position Ref Alt VariantID esp6500siv2_all ExAC_ALL'
' ThousandGenomes_2014oct_all cg46 CADD CADD_Phred Polyphen2_HDIV_score Polyphen2_HDIV_pred'
' Polyphen2_HVAR_score Polyphen2_HVAR_pred consequence Info_Variant1', file=file)
file.close()