Re: [stacks] inbreeding coefficient

[Julian Catchen]

Hi Anna,

Populations does indeed use the equation you specify, from the source code:

//



// Calculate F_is, the inbreeding coefficient of an individual (I)
// relative to the subpopulation (S):
// Fis = (exp_het - obs_het) / exp_het
//



s->nucs[pos].Fis = s->nucs[pos].pi == 0 ? 0 :
(s->nucs[pos].pi - obs_het) / s->nucs[pos].pi;

However, GenoDive may be using a different formulation for it's
calculation of Pi. You can see our formulation for Pi in Appendix S1 of
the 2013 paper.

For average values of population genetics stats within a population, the
populations program produces two tables in the
batch_X.sumstats_summary.tsv file, one that includes the fixed positions
and one that does not. Fis is included in both calculations, but Fis
will always be zero in the case of a fixed position.

Best,

julian



Anna P wrote:
> Hello,
>
> I'm wondering how does the populations program in Stacks calculate the
> average Fis for a population across loci. The values I get don't seem to
> match the equation Fis=pi-H/pi (as described in the supplement of the
> 2013 paper), and I find qualitatively very different values when I use
> other programs (e.g. GenoDive). How does the populations program treat
> the loci that are fixed within a population, when calculating average
> values across loci? To my understanding these loci shouldn't count for
> the calculation of Fis, as both their expected and observed
> heterozygosity is zero, but they should still count when calculating
> genetic diversity statistics, such as pi and observed heterozygosity, is
> that right? I'm using only bi-allelic SNPs.
>
> thank you very much,
>
> Anna
>