Founder population from AlphaSimR

[rthapa1]

Hi , I am simulating a breeding program for synthetic cultivar. I checked the distribution of founder population. The allele frequency distribution looks like skewed inferring fixation of many alleles. I used the following parameter. Could you please see what is happening? I was expecting the allele frequency distribution to be normal distribution.

library(AlphaSimR) trait1=1; nInd=1000; nChr = 8; segSites = 10000; mean.DD.i <- 1 var.DD.i <- 10 h2.i <- 0.1

simulation

FOUNDERPOP = runMacs(nInd=nInd,nChr=nChr,segSites=segSites, inbred=FALSE, ploidy=4)

SP = SimParam$new(FOUNDERPOP) SP = SP$ addTraitAD(60, meanDD=mean.DD.i, varDD=var.DD.i)$ setVarE(h2=h2.i) SIMPARAM = SP$addSnpChip(8000)

founderpop = newPop(FOUNDERPOP, simParam=SIMPARAM) geno <- pullSnpGeno(founderpop)

[gregorgorjanc]

@rthapa1 why would you expect normal distribution for allele frequencies? Most neutral mutations will have low frequency or will be close to fixation by drift, hence U shaped distribution - I am not fully sure what MaCS does, but encoding alleles as 0 or 1 does not make difference for quantitative genetic analyses done in AlphaSimR (hence the half U shape) - we are working on improving this. Also note that the GENERIC species, the default in runMacs(), might not be what you would expect/need for your setting. You can build your own MaCS simulation usingrunMacs2()`.

[rthapa1]

Thank you for your explanation. Is there a way to create population structure in the founder population?

[gregorgorjanc]

@rthapa1 Yes, by creating splits in the coalescent simulation with runMacs() we are working on leveraging other more modern back in time simulators for this as well. There are other posts on this forum that enable you to work with your genotype data, which will have the structure you want, but there are many other issues if you follow that path.

[rthapa1]

Thank you. Is the nChr equals to haploid chromosome number or basic chromosome number? I am working on autotetraploid so I assume it would be 8. I would appreciate your answer.

[gregorgorjanc]

Basic chromosome number. We do have support for autotetraploids - see ploidy argument in runMacs()