Feature request: print read ids of reads supporting variant

[scottgigante]

Is it possible to dump the read ids corresponding to those reads which supported a variant from nanopolish variants? I would like to use this for haplotype phasing.

[jts]

One possible workflow is to call variants with nanopolish variants, then calculate read support using phase-reads. That is a bit roundabout though.

If I add this to nanopolish variants what output format would you want? I could put it in the VCF INFO tags but the lines would be very long.

[scottgigante]

I did play around with phase-reads but even then this only calls SNPs, where preferably I would have everything.

I agree it's probably a bit long to include in the VCF file, even if it's optional. Would it make more sense to optionally spit out another file in a similar form to what call-methylation does?

chromosome      pos     read_name       log_lik_ratio   log_lik_ref      log_lik_alt    REF .    ALT

(I assume that you're calculating a log likelihood for each read, thresholding, and then calculating % support from that - correct me if I'm wrong)