Mercurial > repos > nick > allele_counts

view tests/artificial.vcf.in @ 10:7f19e8c03358 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload for repository https://github.com/galaxyproject/dunovo commit 19875a8a45eb5e40a47fe177deafe690fb4f04fb"
author nick
date Tue, 31 Mar 2020 20:24:27 -0400
parents 31361191d2d2
children
line wrap: on
line source

##fileformat=VCFv4.1
##comment="ARGS=-r 1 -f 10 -c 10"
##comment="This is a test set of made-up sites, each created in order to test certain functionality. It's meant to be run with -f 10 -c 10"
##fileDate=19700101
##source=Dan
##reference=file:///scratch/dan/galaxy/galaxy-central/database/files/002/dataset_0000.dat
##INFO=<ID=AC,Number=A,Type=Integer,Description="Allele count in genotypes, for each ALT allele, in the same order as listed">
##INFO=<ID=AF,Number=A,Type=Float,Description="Allele Frequency, for each ALT allele, in the same order as listed">
##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">
##FORMAT=<ID=AC,Number=.,Type=Integer,Description="Allele count in genotypes, for each ALT allele, in the same order as listed">
##FORMAT=<ID=AF,Number=.,Type=Float,Description="Allele Frequency, for each ALT allele, in the same order as listed">
##FORMAT=<ID=NC,Number=.,Type=String,Description="Nucleotide and indel counts">
#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO	FORMAT	THYROID
# General note: the only data made consistent is the CHROM, POS, REF, ALT, and the variant data (after the ':'). The other stuff isn't supposed to be consistent.
# Simplest case, but POS 0 and no minor allele
chr1	0	.	A	.	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=15,-A=15,
# Simple, normal cases of A/G variants: above/below threshold x strand bias/no strand bias (2 x 2 = 4 cases)
chr1	10	.	A	G	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=15,+G=1,-A=15,-G=1,
chr1	20	.	A	G	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=15,+G=1,-A=16,
chr1	30	.	A	G	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=10,+G=2,-A=11,-G=2,
chr1	40	.	A	G	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=11,+G=3,-A=11,
# Same 4 cases, but with minor allele = N
chr1	50	.	A	N	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=15,+N=1,-A=15,-N=1,
chr1	60	.	A	N	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=15,+N=1,-A=16,
chr1	70	.	A	N	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=10,+N=2,-A=11,-N=2,
chr1	80	.	A	N	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=11,+N=3,-A=11,
# Same 4 cases, but with an additional noncanonical minor allele d1
chr1	82	.	A	d1	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=15,+G=1,+d1=1,-A=15,-G=1,-d1=1,
chr1	84	.	A	d1	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=15,+G=1,+d1=1,-A=16,
chr1	86	.	A	d1	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=10,+G=2,+d1=1,-A=11,-G=2,-d1=1,
chr1	88	.	A	d1	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=11,+G=3,+d1=1,-A=11,
# Same 4 cases, but with minor allele = d1 (non-canonical)
chr1	90	.	A	d1	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=15,+d1=1,-A=15,-d1=1,
chr1	100	.	A	d1	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=15,+d1=1,-A=16,
chr1	110	.	A	d1	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=10,+d1=2,-A=11,-d1=2,
chr1	120	.	A	d1	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=11,+d1=3,-A=11,
# Same 4 cases, but with MAJOR allele = d1
chr1	130	.	d1	G	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+d1=15,+G=1,-d1=15,-G=1,
chr1	140	.	d1	G	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+d1=15,+G=1,-d1=16,
chr1	150	.	d1	G	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+d1=10,+G=2,-d1=11,-G=2,
chr1	160	.	d1	G	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+d1=11,+G=3,-d1=11,
# Test edge cases where freq == freq_thres and/or covg == covg_thres
chr1	200	.	A	.	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=9,-A=9,
chr1	210	.	A	.	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=10,-A=10,
chr1	220	.	A	.	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=11,-A=11,
chr1	230	.	A	G	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=91,+G=9,-A=91,-G=9,
chr1	240	.	A	G	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=90,+G=10,-A=90,-G=10,
chr1	250	.	A	G	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=89,+G=11,-A=89,-G=11,
# Test case where minor allele is above threshold on only one strand because of different coverage. Also, a long decimal minor allele frequency.
chr1	260	.	A	G	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=13,+G=7,-A=93,-G=7,
# Test case where minor alleles have equal frequency: Above/below threshold, +/- strand bias
chr1	300	.	T	G,A	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=1,+G=1,+T=38,-A=1,-G=1,-T=38,
chr1	310	.	T	G,A	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=6,+G=6,+T=38,-A=6,-G=6,-T=38,
chr1	320	.	T	G,A	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=6,+G=6,+T=38,-A=6,-G=6,-T=18,
chr1	330	.	T	G,A	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=6,+G=6,+T=38,-A=1,-G=1,-T=28,
chr1	340	.	T	G,A	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=1,+T=80,-G=1,-T=18,
chr1	350	.	T	G,A	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=11,+T=60,-G=11,-T=18,
# Case where + and - variants are interleaved with each other. Also, a long decimal result for the minor allele frequency.
chr1	400	.	A	N	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=16,-A=16,+G=4,-G=4,
# Test complex data in the ALT, INFO, and sample (before the ':') columns
chr1	410	.	T	G,A	.	.	AC=1,1;AF=0.0111111111111,0.0111111111111	GT:AC:AF:NC	0/0:1,1:0.0111111111111,0.0111111111111:+A=1,+T=81,-T=16,-G=2,
chr1	420	.	A	.	.	.	AC=;AF=	GT:AC:AF:NC	0/0:::+A=82,-A=22,
# Test some other types of noncanonical variants (tie for 2nd and not)
chr1	430	.	A	N,GAA,d2	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=15,+N=1,+d2=1,-A=15,-GAA=2
chr1	440	.	A	N,GAA,d2	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=15,+N=1,+d2=1,-A=15,-GAA=1
# No canonical variants present
chr1	450	.	A	d1	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+d1=15,-d1=20,
chr1	460	.	A	d1,N	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+d1=15,+N=2,-d1=20,-N=2,
# Catch some divide by zero errors
chr1	470	.	A	.	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=0,
# Test an unusual CHROM value and a long POS value
27	1234567890	.	A	N	.	.	AC=;AF=	GT:AC:AF:NC	0/0:1:1:+A=15,+N=1,-A=14,