Petr Danecek

This happens when the reference alleles don't match across files. Unfortunately, your description does not give enough information, it's hard to say what is going on without seeing the data....

Possibly, it depends how big are the mandatory columns (CHROM-INFO) compared to the FORMAT fields. Why don't you look in the output file and compare it with the input file?...

If the spanning star allele is the only ALT, then simply removing the record would work. However, not for cases like this ``` #CHROM POS REF ALT GT chr1 154590148...

What version of bcftools are you using? I'd try to update to the latest and see if it changes anything. If not, a small test case would be necessary to...

There seems to be some confusion. Running with `--local-csq` or without can give different outputs, and that's a feature, not a bug. In the haplotype-aware mode multiple variants can be...

Check if there are any reads in that region ``` samtools view your.bam chrX:start-end | less ``` A common problem is a mismatch in chromosome naming convention (`chrX` vs `X`)...

The next thing to check then is what raw mpileup looks like. Maybe all positions in that region are non-variant, therefore `bcftool call -v` removes everything? You can test with...

In the first screenshot all bases in all reads are identical to the reference, hence there are no variants, nothing to call. The second screenshot shows several bases with no...

This requires a test case to reproduce and debug the problem. This script can be used to create a small slice of the bam and the reference https://github.com/pd3/mpileup-tests/blob/main/misc/create-bam-test

> It's unclear how we can use this as a conversion tool for an existing file. I can see two possible routes. > > 1. Take an existing many-sample VCF...