Petr Danecek

That sounds very odd indeed. This is such a basic functionality that most likely you are doing something wrong. I am unable to help without a small test case.

This should work `--targets-file ^file.txt`, see the documentation here http://samtools.github.io/bcftools/bcftools.html Please open a new issue next time, this is unrelated to the problem discussed here

The program has a limitation, when a VCF is used as the source of annotations, it can match a line only once. You'd have to split the multiallelic records into...

This looks like the IUPAC ambiguity code. I suspect this comes from your reference file, bcftools prints only what it encounters. Can you check the output of ``` samtools faidx...

It's a bit confusing, isn't it. The option `--multi-overlaps` is tied to the `--multiallelics` option. This one splits multiallelic sites into biallelic sites, for example the row `AA > A,CC`...

Variants are normalized after splitting. (Note there were some improvements in the latest release 1.20 to handle simultaneous atomization and splitting of multiallelics.) Regarding 'both' and 'any', it is described...

The FILTER column was intended as an indicator that a site is truly polymorphic, rather than making assertions about the alleles. The meaning would be changed by this, all the...

The meaning of a single dot is ambiguous, it can mean both.

The missing genotype was introduced by the VCF specification specifically for cases with missing data: in the absence of data (./.) we have no supporting evidence for the claim that...

> Given GT is conditional on a variant, this means that. GT is meaningless for 0/0 calls. It's, by definition 0 and any other value is violates either VCF specifications...