Help with Simulating Beneficial SNP and Linked Neutral Mutations

[Stylianos Mavrianos]

Hi SLiM Community,

I'm new to using SLiM and could use some help with a simulation. I'm trying to simulate a beneficial SNP with neutral mutations hitchhiking within a 500 base pair region on either side. However, I'm running into some issues.

What I Tried:

• Used a rate map to set mutation rates and identify the beneficial SNP.

Issue:

• Random neutral mutations are hitchhiking with the SNP, but the ones in the 500 base pair flanking region are not.

Questions:

1. How can I ensure the neutral mutations in the flanking regions hitchhike with the beneficial SNP?
2. Are there specific parameters or settings I should adjust in the rate map or elsewhere?

Here's my current script and rate map configuration:

Thank you.

[Ben Haller]

Hi Stylianos!

There's a lot going on in your model.  You've got a fairly complicated recombination rate map; I'm not sure what the 1e-90 recombination rate points are meant to represent, and then you've got a pattern of low and high recombination rates along the genome.  If that's the specific scenario you want to model, then that's fine.  I'd just note that hitchhiking is observed even with a uniform recombination rate along the genome; it is not necessarily a consequence of a lower recombination rate surrounding the point where a mutation occurs.  I just thought I'd mention this, in case the model you're building is based upon a misunderstanding of how/why hitchhiking happens.

So we could make the recombination rate map just be a single uniform rate, and still ask: why is hitchhiking not strongly observed in this model?  I'd suggest that a likely reason is that you introduce three beneficial mutations, at three positions widely spaced along the genome, in the same single genome object in the same tick.  This means that all of the final generation, when the sweep of all three mutations fixes, will descend from that same single ancestral genome at all three of those positions.  You might get a stray recombination near one of the ends of the genome that bring in a little ancestry from somewhere else at an end; and you might get a pair of recombinations in the middle somewhere that bring in a chunk of ancestry from somewhere else in the middle.  But the three beneficial mutations in the same initial genome are helping each other sweep, through linkage effects, and in the interior a double recombination is needed to bring in a chunk of ancestry from elsewhere, so the hitchhiking pattern will probably be less pronounced than it would be with just a single sweep mutation (i.e., most/all of the genome will tend to hitchhike, rather than just a region around a single sweep mutation).  In support of this explanation, if I turn off the outer two beneficials and just have the central beneficial sweep, then I see a clear pattern of hitchhiking around it, even in a single run.

But I'd also say that I *do* see hitchhiking when I run your model, fairly often.  To see it better, it's helpful to crank up the mutation rate to, say, 1e-5, and turn on display of substitutions in SLiMgui (since the mutations that hitchhike all the way from the original sweep ancestor will fix, along with the beneficial mutations, and will thus be converted into substitution objects).  With those settings, I've attached a screenshot from an example run I just did. 



What you can see is that substitutions in the neutral mutations (indicating complete hitchhiking) are present across much of the genome, but that in two chunks substitutions have *not* occurred; instead we see blocks of yellow mutations that are still segregating (indicating incomplete hitchhiking, where some diversity from an ancestor other than the original sweep ancestor is still present in the simulation due to a recombination at some point during the sweep).  Those chunks of incomplete hitchhiking are indeed in areas that do not include the beneficial mutations themselves (as they must be).  This doesn't happen every run; in many runs, just a bit of edge recombines, or there is no recombination at all and the whole genome sweeps; but patterns like the one I snapped do occur in a minority of runs.  So I think your model is, in fact, working, but perhaps the substitution process obscured the result for you, or the mutation rate you were using was so low that the hitchhiking pattern was not immediately obvious.

I hope this is helpful.  Happy modeling!

Cheers,
-B.

Benjamin C. Haller
Messer Lab
Cornell University


Stylianos Mavrianos wrote on 7/17/24 6:18 AM:

--
SLiM forward genetic simulation: http://messerlab.org/slim/
---
You received this message because you are subscribed to the Google Groups "slim-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to slim-discuss...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/slim-discuss/16f79b9c-210b-4cbd-9f56-483865e14f9en%40googlegroups.com.

[Stylianos Mavrianos]

Hi Ben,

Thank you very much for your  quick answer, I will look into it.

Best,

Stylianos