Hi all,
I am hoping for a sanity check on RoH (bcftools) so I can feel confident in my analysis since I am using an atypical dataset but I think the tool is appropriate.
I have nonhuman primate exome data (since exome capture works well for primates) and it is aligned to a congener reference assembly. I was worried about analyzing samples that were aligned to a different species because it seems like that could produce overestimates of homozygosity at sites that are fixed differences between species (would be considered homozygous rather than invariant). But I think this is fine since I have ~30 individuals from the same populations so I can estimate allele frequencies from my data so I am assuming that fixed differences between species will not lead to overestimates of homozygosity.
In my results, I am getting RoHs that are longer than what is continuously represented in the exome (and in fact, all individuals of sums of ROH greater than the total size of the genome captured). This was confusing at first. But my understanding is that this is what is supposed to happen - that the exome data are essentially being used as marker data and that the HMM takes into account their discontinuousness and sparsity...it's just a little weird to wrap my head around.
Thank you for any thoughts on whether bcftools roh would be appropriate for analyzing nonmodel exome data that has been aligned to a different species :) . I need help confirming what I am doing/assuming is not wrong. Bcftools roh is a great tool and how to implement it is nicely documented, but how it works (I've read the original pub) is not extensively broken down. This is an endangered species, so this is meant as a conservation genetics analysis (looking at inbreeding, possible bottleneck, and potentially inbreeding depression as enrichment of deleterious variants in ROH). Also, I am using the human recombination rate, which may be an overestimate, but probably the best choice.
Thank you!!
