The VariantsPerSample annotator is invoked through the SVVariantAnnotator walker, which defines arguments common to all annotators.
The VariantsPerSample annotator is a simple annotator that counts how many variants are in each sampled genome. The output can be rolled up into population level statistics.
In the current implementation, VariantsPerSample uses the
GSSAMPLES INFO tag in the input VCF file to determine which samples carry the variant. The input VCF file should not contain genotypes.
Filtered variants are not counted in the totals.
-populationMap <map-file>: A tab-delimited input file containing two columns: the sample ID and a population ID for that sample. If supplied, the population information will be carried over into the output report.
No VCF annotations are produced, but this annotator is used to produce a report file. The report file will contain one line per sample. The report includes the number of variants and also the population if -populationMap is supplied.
java -Xmx2g -cp SVToolkit.jar:GenomeAnalysisTK.jar \ org.broadinstitute.sting.gatk.CommandLineGATK \ -T SVVariantAnnotator \ -A VariantsPerSample \ -R /humgen/1kg/reference/human_g1k_v37.fasta \ -BTI variant \ -B:variant,VCF input.vcf \ -populationMap sample_to_population.map \ -writeReport \ -reportFile variants_per_sample.dat