Monomorphism from simulation

[Zuxi Cui]

Hi there,

I'm experiencing an issue with SLiM creating weird monomorphism via simulation.

I'm using the following example to explain the problem.

I input variants in VCF format to the SLiM and simulate offspring using WF model with 30 mating generations. One typical variant is shown in the screenshot below.

After that simulation, SLiM output this variant as the following:

The variant became monomorphic with a change in REF allele.

The sample size of the input cohort is 90, and the sample size of each generation and the output cohort is 1,300.

I observed a great amount (>30% of all variants) of this type of change to monomorphic.

This seems abnormal even by genetic drift, as the allele frequencies of these input variants are not rare. In addition, samples are not coded as regular homozygous (1|1 or 0|0) in the output. The only thing that determines the monomorphism is the change of REF/ALT to make REF=ALT.

Is there a way to test if these monomorphic sites are indeed generated by genetic drift or any type of coding error?

I've included the code in the attachment. Let me know if you have any comments.

Best,

Terry

[Ben Haller]

Hi Terry,

It's hard to guess.  I guess I'd suggest perhaps saving out a VCF file at each generation, and investigating what happens at this position from generation to generation (using the same random number seed to get reproducible behavior).  It does seem odd, but I'm not sure what's happening.  I'd just gather more evidence and investigate what the process looks like, at this stage.  If it's a bug, presumably the pattern over time would look quite different from if it is drift.

Cheers,
-B.

Benjamin C. Haller
Messer Lab
Cornell University