Allele size confidence intervals (feature request)

[hdashnow]

Have you considered adding allele size confidence intervals?
Or even min and max size of allele in supporting reads would be a useful proxy.

cc'ing @elbayaliyev

[sowpati]

Hi @hdashnow and @elbayaliyev

Thank you for raising this feature request. As it happens, just a few hours before this, we pushed a newer version (0.3.1) to dev, which adds the allele range for each haplotype to the VCF output 😁

This is added to the FORMAT field with 'AR' tag. For now, we are reporting the min and max of supporting reads for each allele (and just repeated twice for homozygous loci for the sake of consistency in parsing). But, we are in the process of adding 95% CI calculations as well.

It would be great if you can test out the dev version (some bugs may be expected) and see how it performs in your case, and we would love feedback/suggestions.

[hdashnow]

Well, that is excellent timing! We will test it out.
My first thought is that if the reads are phased, the AR tag should be different for each allele, even when homozygous for the allele size/sequence.

[sowpati]

Thanks Harriet. Your suggestion of different ARs for phased homozygous alleles makes sense. We'll be implementing this in the upcoming version.

[sowpati]

Okay, turns out - we do report two different ARs for homozygous loci if the locus is phased. However, many loci which are unambiguously homozygous don't undergo the phasing process. For such loci, we are reporting two ARs, but one of them is a dummy, where the inferred allele length is reported as the range. We will update the documentation to highlight these aspects, but we plan to revisit this part as well in the new version.

[elbayaliyev]

I can confirm that i succesfully ran 3.0.1 and i can see Allele length range tag. Thanks !!!