Probes for SNPs/indels should not extend beyond the variant - all kmers should overlap the variant. Otherwise, you add coverage to both alleles from shared flanks. Not a massive deal for illumina (although why add pointless noise) but I think would be a bigger issue for nanopore data when we already have enough problems without adding to them.
That said, I'm sure I have seen/mostly see things where one allele has covg 0.
So probes can't be always 63, as Phelim speculated today.
Phelimb
commented
5 years ago
Probes for SNPs/indels should not extend beyond the variant - all kmers should overlap the variant. Otherwise, you add coverage to both alleles from shared flanks. Not a massive deal for illumina (although why add pointless noise) but I think would be a bigger issue for nanopore data when we already have enough problems without adding to them.
That said, I'm sure I have seen/mostly see things where one allele has covg 0. So probes can't be always 63, as Phelim speculated today.
ie maybe there is no issue here