def filter_snp(general, snp_info, sample_stats, mincov, minsnp, ploidy):
ref = general[2]
dp4 = map(int,
snp_info.get('DP4', '0,0,0,0').split(
',')) # fw.ref, rev.ref, fw.alt, rev.alt (filtered)
total_reads = sum(dp4)
if total_reads < mincov: # coverage low
return '0' # '/'.join([ref]*ploidy)
ratio = 100.0 * (dp4[2] + dp4[3]) / total_reads
if ratio < minsnp / ploidy:
return "-"
alt = general[3]
alts = [ref] + alt.split(',')
sample = sample_stats[1] # ex: "0/1:48,0,53:14:0:50"
genotype = sample.split(':')[
0] # todo: better according to *format* = GT:PL:DP:SP:GQ
sep = '/' if '/' in genotype else '|' # phased if |, unphased if /
genotype = [alts[int(i)] for i in genotype.split(sep)]
# Diploid: GT: '0/1' -> 'T/A (50% of 8)' [if ref='T' and alt='A']
# Haploid: GT: '0/1' -> 'A (50% of 8)'
if ploidy == 1:
genotype = unique(genotype)
if ref in genotype: genotype.remove(ref)
genotype = sep.join(genotype)
genotype = "%s (%.0f%% of %d)" % (genotype, ratio, total_reads)
return genotype
def _write_buffer(_buffer, outex):
new_codon = None
# One position at a time
for chr, pos, refbase, variants, cds, strand, ref_codon, shift in _buffer:
varbase = list(
variants
) # Ex: ['G','G','G','T (56% of 167)','G'], ['A/A','G/G (100% of 7)']
if new_codon is None:
new_codon = [[ref_codon] for _ in range(len(varbase))]
variants = [] # [[variants sample1], [variants sample2], ...]
# One sample at a time
for variant in varbase:
if variant in ['0', '-']:
variants.append([refbase])
else: # Ex: 'C/G (80% of 10)' : heterozygous simple (ref is C) or double snp (ref is not C)
v = variant.split()[0] # Ex: C/G
v = unique(v.split('/')) # Ex: G/G -> 'G'
if refbase in v:
v.remove(refbase) # Ex: C/G -> 'G' (if ref is C)
variants.append(v)
# One sample at a time
for k, v in enumerate(variants):
cnumb = len(new_codon[k])
newc = new_codon[k] * len(v)
for i, vari in enumerate(v):
for j in range(cnumb):
newc[i * cnumb +
j] = newc[i * cnumb +
j][:shift] + vari + newc[i * cnumb +
j][shift + 1:]
assert ref_codon[
shift] == refbase, "bug with shift within codon"
new_codon[k] = newc
if new_codon is None: return
if strand == -1:
ref_codon = revcomp(ref_codon)
new_codon = [[revcomp(s) for s in c] for c in new_codon]
for chr, pos, refbase, variants, cds, strand, dummy, shift in _buffer:
refc = [iupac.get(x, x) for x in ref_codon]
ref_codon = [''.join(x) for x in product(*refc)]
newc = [[[iupac.get(x, x) for x in variant] for variant in sample]
for sample in new_codon]
new_codon = [[
''.join(x) for codon in sample for x in product(*codon)
] for sample in newc]
if refbase == "*":
result = [chr, pos+1, refbase] + list(variants) + [cds, strand] \
+ [','.join([translate.get(refc,'?') for refc in ref_codon])] + ["indel"]
else:
result = [chr, pos+1, refbase] + list(variants) + [cds, strand] \
+ [','.join([translate.get(refc,'?') for refc in ref_codon])] \
+ [','.join([translate.get(s,'?') for s in newc]) for newc in new_codon]
outex.write("\t".join([str(r) for r in result]) + "\n")
def _write_buffer(_buffer, outex):
new_codon = None
# One position at a time
for chr,pos,refbase,variants,cds,strand,ref_codon,shift in _buffer:
varbase = list(variants) # Ex: ['G','G','G','T (56% of 167)','G'], ['A/A','G/G (100% of 7)']
if new_codon is None:
new_codon = [[ref_codon] for _ in range(len(varbase))]
variants = [] # [[variants sample1], [variants sample2], ...]
# One sample at a time
for variant in varbase:
if variant in ['0','-']:
variants.append([refbase])
else: # Ex: 'C/G (80% of 10)' : heterozygous simple (ref is C) or double snp (ref is not C)
v = variant.split()[0] # Ex: C/G
v = unique(v.split('/')) # Ex: G/G -> 'G'
if refbase in v: v.remove(refbase) # Ex: C/G -> 'G' (if ref is C)
variants.append(v)
# One sample at a time
for k,v in enumerate(variants):
cnumb = len(new_codon[k])
newc = new_codon[k]*len(v)
for i,vari in enumerate(v):
for j in range(cnumb):
newc[i*cnumb+j] = newc[i*cnumb+j][:shift] +vari +newc[i*cnumb+j][shift+1:]
assert ref_codon[shift] == refbase, "bug with shift within codon"
new_codon[k] = newc
if new_codon is None: return
if strand == -1:
ref_codon = revcomp(ref_codon)
new_codon = [[revcomp(s) for s in c] for c in new_codon]
for chr,pos,refbase,variants,cds,strand,dummy,shift in _buffer:
refc = [iupac.get(x,x) for x in ref_codon]
ref_codon = [''.join(x) for x in product(*refc)]
newc = [[[iupac.get(x,x) for x in variant] for variant in sample]
for sample in new_codon]
new_codon = [[''.join(x) for codon in sample for x in product(*codon)] for sample in newc]
if refbase == "*":
result = [chr, pos+1, refbase] + list(variants) + [cds, strand] \
+ [','.join([translate.get(refc,'?') for refc in ref_codon])] + ["indel"]
else:
result = [chr, pos+1, refbase] + list(variants) + [cds, strand] \
+ [','.join([translate.get(refc,'?') for refc in ref_codon])] \
+ [','.join([translate.get(s,'?') for s in newc]) for newc in new_codon]
outex.write("\t".join([str(r) for r in result])+"\n")
def filter_snp(general,snp_info,sample_stats,mincov,minsnp,ploidy):
ref = general[2]
dp4 = map(int, snp_info.get('DP4','0,0,0,0').split(',')) # fw.ref, rev.ref, fw.alt, rev.alt (filtered)
total_reads = sum(dp4)
if dp4[2]+dp4[3] < mincov: # too few supporting alt
return '-' # '/'.join([ref]*ploidy)
ratio = 100.0*(dp4[2]+dp4[3])/total_reads
if ratio < minsnp/ploidy:
return "0"
alt = general[3]
alts = [ref]+alt.split(',')
sample = sample_stats[1] # ex: "0/1:48,0,53:14:0:50"
genotype = sample.split(':')[0] # todo: better according to *format* = GT:PL:DP:SP:GQ
sep = '/' if '/' in genotype else '|' # phased if |, unphased if /
genotype = [alts[int(i)] for i in genotype.split(sep)]
# Diploid: GT: '0/1' -> 'T/A (50% of 8)' [if ref='T' and alt='A']
# Haploid: GT: '0/1' -> 'A (50% of 8)'
if ploidy == 1:
genotype = unique(genotype)
if ref in genotype: genotype.remove(ref)
genotype = sep.join(genotype)
genotype = "%s (%.0f%% of %d)" % (genotype,ratio,total_reads)
return genotype