How to filter INFO field of VCF by 'OR' using bcftools

Question

[adixon3]

I have a VCF from dbVar (https://ftp.ncbi.nlm.nih.gov/pub/dbVar/data/Homo_sapiens/by_study/vcf/nstd102.GRCh38.variant_call.vcf.gz) and I want to pull out all the variants which have ANY the following values in the CLNSIG field: Pathogenic, Likely_pathogenic.

If I use the following command, bcftools delivers only variants with "Pathogenic" in the CLNSIG field:

bcftools view -i 'INFO/CLNSIG ~ "Pathogenic"|"Likely_pathogenic"' nstd102.GRCh38.variant_call.vcf.gz

How do I properly use the 'OR' operator to find the union of variants that match my list of CLNSIG values?

Answer 1

[Pierre Lindenbaum]

try

    bcftools view -i 'INFO/CLNSIG ~ "Pathogenic\|Likely_pathogenic"' nstd102.GRCh38.variant_call.vcf.gz

[adixon3]

Unfortunately this produces a VCF with a header and no variants :( I've noticed that this also happens with the following command

##fileformat=VCFv4.1
##FILTER=<ID=PASS,Description="All filters passed">
##fileDate=20231012
##reference=GCF_000001405.40
##ALT=<ID=<CNV>,Description="Copy number variable region">
##ALT=<ID=<DEL>,Description="Deletion relative to the reference">
##ALT=<ID=<DUP>,Description="Region of elevated copy number relative to the reference">
##ALT=<ID=<INS>,Description="Insertion of sequence relative to the reference">
##ALT=<ID=<INV>,Description="Inversion of reference sequence">
##INFO=<ID=DBVARID,Number=1,Type=String,Description="ID of this element in dbVar">
##INFO=<ID=CIEND,Number=2,Type=Integer,Description="Confidence interval around END for imprecise variants">
##INFO=<ID=CIPOS,Number=2,Type=Integer,Description="Confidence interval around POS for imprecise variants">
##INFO=<ID=DESC,Number=1,Type=String,Description="Any additional information about this call (free text, enclose in double quotes)">
##INFO=<ID=END,Number=1,Type=Integer,Description="End position of the variant described in this record">
##INFO=<ID=IMPRECISE,Number=0,Type=Flag,Description="Imprecise structural variation">
##INFO=<ID=SVTYPE,Number=1,Type=String,Description="Type of structural variant">
##INFO=<ID=SVLEN,Number=.,Type=String,Description="Difference in length between REF and ALT alleles">
##INFO=<ID=CHR2,Number=1,Type=String,Description="Second (To) Chromosome in a translocation pair">
##INFO=<ID=REGIONID,Number=.,Type=String,Description="The parent variant region accession(s)">
##INFO=<ID=EXPERIMENT,Number=1,Type=Integer,Description="The experiment_id (from EXPERIMENTS tab) of the experiment that was used to generate this call">
##INFO=<ID=EVENT,Number=.,Type=String,Description="The parent variant region accession of a mutation event">
##INFO=<ID=LINKS,Number=.,Type=String,Description="Link(s) to external database(s) - see LINKS tab of dbVar submission template for examples">
##INFO=<ID=CLNSIG,Number=.,Type=String,Description="Clinical significance for this single variant">
##INFO=<ID=CLNACC,Number=.,Type=String,Description="Accessions and version numbers assigned by ClinVar">
##INFO=<ID=clinical_source,Number=1,Type=String,Description="Source of clinical significance">
##INFO=<ID=SOMATIC,Number=0,Type=Flag,Description="Indicates that the record is a somatic mutation. NOT for clinical assertions, i.e. cancer. See also ORIGIN.">
##INFO=<ID=ORIGIN,Number=1,Type=String,Description="Origin of allele, if known; should be one of (biparental, de novo, germline, inherited, maternal, not applicable, not provided, not-reported, paternal, tested-inconclusive, uniparental, unknown, see ClinVar for details). See also SOMATIC">
##INFO=<ID=PHENO,Number=.,Type=String,Description="Phenotype(s) thought to associated with this call. NOT for clinical assertions (submit to ClinVar). (free text, enclose in double quotes)">
##INFO=<ID=SAMPLE,Number=1,Type=String,Description="sample_id from dbVar submission; every call must have SAMPLE or SAMPLESET, but NOT BOTH">
##INFO=<ID=SAMPLESET,Number=1,Type=Integer,Description="sampleset_id from dbVar submission; every call must have SAMPLESET or SAMPLE but NOT BOTH">
##INFO=<ID=VALIDATED,Number=0,Type=Flag,Description="Validated by follow-up experiment">
##INFO=<ID=SEQ,Number=1,Type=String,Description="Variation sequence">
##INFO=<ID=AC,Number=.,Type=Integer,Description="Global Allele count">
##INFO=<ID=AF,Number=.,Type=Float,Description="Global Allele frequency">
##INFO=<ID=AN,Number=.,Type=String,Description="Global Allele name">
##contig=<ID=1>
##contig=<ID=2>
##contig=<ID=3>
##contig=<ID=4>
##contig=<ID=5>
##contig=<ID=6>
##contig=<ID=7>
##contig=<ID=8>
##contig=<ID=9>
##contig=<ID=10>
##contig=<ID=11>
##contig=<ID=12>
##contig=<ID=13>
##contig=<ID=14>
##contig=<ID=15>
##contig=<ID=16>
##contig=<ID=17>
##contig=<ID=18>
##contig=<ID=19>
##contig=<ID=20>
##contig=<ID=21>
##contig=<ID=22>
##contig=<ID=MT>
##contig=<ID=NT_113793.3>
##contig=<ID=NT_113796.3>
##contig=<ID=NT_187361.1>
##contig=<ID=NT_187513.1>
##contig=<ID=NT_187562.1>
##contig=<ID=NT_187576.1>
##contig=<ID=NT_187593.1>
##contig=<ID=NT_187594.1>
##contig=<ID=NT_187600.1>
##contig=<ID=NT_187603.1>
##contig=<ID=NT_187606.1>
##contig=<ID=NT_187613.1>
##contig=<ID=NT_187614.1>
##contig=<ID=NT_187620.1>
##contig=<ID=NT_187633.1>
##contig=<ID=NT_187648.1>
##contig=<ID=NT_187653.1>
##contig=<ID=NT_187660.1>
##contig=<ID=NT_187661.1>
##contig=<ID=NT_187681.1>
##contig=<ID=NT_187682.1>
##contig=<ID=NT_187693.1>
##contig=<ID=NW_003571049.1>
##contig=<ID=NW_003571054.1>
##contig=<ID=NW_003571055.2>
##contig=<ID=NW_003571056.2>
##contig=<ID=NW_003571057.2>
##contig=<ID=NW_003571058.2>
##contig=<ID=NW_003571060.1>
##contig=<ID=NW_003571061.2>
##contig=<ID=NW_009646195.1>
##contig=<ID=NW_009646198.1>
##contig=<ID=NW_009646206.1>
##contig=<ID=NW_009646209.1>
##contig=<ID=NW_011332698.1>
##contig=<ID=NW_011332701.1>
##contig=<ID=NW_012132918.1>
##contig=<ID=NW_015148966.1>
##contig=<ID=NW_015495298.1>
##contig=<ID=NW_018654714.1>
##contig=<ID=X>
##contig=<ID=Y>
##bcftools_viewVersion=1.13+htslib-1.13
##bcftools_viewCommand=view -i 'INFO/CLNSIG ~ "Pathogenic\|Likely_pathogenic"' nstd102.GRCh38.variant_call.vcf.gz; Date=Mon Jan  6 09:00:06 2025
#CHROM  POS ID  REF ALT QUAL    FILTER  INFO

[Pierre Lindenbaum]

##bcftools_viewVersion=1.13+htslib-1.13

your version of bcftools is just too old. https://github.com/samtools/bcftools/releases/tag/1.13

[adixon3]

Unfortunately, bcftools 1.21 produces the same results. Only thing that actually filters is a single filter term using either of the two styles:

bcftools view -i 'INFO/CLNSIG ~ "Pathogenic"' nstd102.GRCh38.variant_call.vcf.gz

or

bcftools view -i 'CLNSIG="Pathogenic"' nstd102.GRCh38.variant_call.vcf.gz

As soon as I add an OR operator (e.g. "type|type_2"), it just outputs a header-only VCF.

[Pierre Lindenbaum]

Unfortunately, bcftools 1.21 produces the same results.

that's strange. that syntax works on my machine.

wget -O - "https://github.com/lindenb/jvarkit/raw/refs/heads/master/src/test/resources/gnomad.genomes.r2.0.1.sites.1.vcf.gz" | bcftools view -i 'INFO/VQSR_culprit  ~ "MQ\|FS" '