Hi everyone,
I have some GBS data from some populations of stoats, and would love some help interpreting the inbreeding coefficients from PLINK.
I have attached all of them, which I graphed per population.
Below are the outputs for --het and --ibc, all theoretically inbreeding coefficients.
This is GBS data, filtered on minor allele frequency and coverage so it's got good coverage and rare alleles are gone. ~6000 SNPs. I have some biological knowledge of the populations and feel pretty confident I know the relative diversity of each. The genome quality is excellent.
Den is Denmark, and probably in reality has the highest genetic diversity. UK likely has the next highest diversity ( these populations are both within the native range). Ire is Ireland, and with one individual here which we know is diverged from all others I think we will have lost most of its diversity in filtering. All other pops are introduced range (NZ). ARK and HUN are near each other, & should be about the same. TAR probably has the most diversity of the NZ samples. RAN is a moderately isolated island near ARK and HUN. WAI is a very isolated highly inbred population on an island.
I have diversity values for all of these based on previous micro satellite work, and HUN and ARK are about the same diversity, RAN is lower, and WAI is much lower.
It looks in most plots to me, unless I'm confused by Plinks outputs, that DEN is most inbred, followed by IRE and UK. If higher Fhat1 was less inbred, that plot would be what I would expect, but I think its the opposite? This could be ascertainment bias due to NZ dominating the sampling?
UK is most like the NZ ones because NZ stoats came from UK - therefore it's most likely I have retained SNPs in UK because my dataset is dominated by NZ variation. WAI should always have higher inbreeding than the other NZ populations I'm sure.
While WAI was a large proportion of the data, it's not more than half, so I don't think ascertainment bias should affect it that much. If as I said, fhat1 is reversed and higher = less inbred, lower = more, than we are all groovy. If not, what do you think caused this?
My guess is that possibly F is the most accurate here and useful, and it is an ascertainment bias for DEN, IRE that draw them up. There may be some sample quality issues with TAR adding some outliers with high inbreeding coefficients.
Other options are that the coverage stats are off and this has caused some funny results?
Any help greatly appreciated!
--ibc calculates inbreeding coefficients for each sample and get the report as
--out plink.ibc
You are going to get different outcomes for inbreeding: ibd of all pairs, and genome wide inbreeding of all pairs based on ibs. You should be able to extract the particular outcome you want.
- Badges
- Science topic
- Similar topics
- Cognitive Science
- Thinking