Using ploidy option versus number of individuals * ploidy

[isinaltinkaya]

Hi,

I have run bunch of simulations aiming to simulate diploid individuals using number_of_individuals * 2 as sample size. I just noticed that there is also a ploidy option. Are there any difference between using ploidy=2 and ploidy=default; sample_size=size*2 to simulate diploid samples? I would assume not since these would be haplotype counts, but I could not be sure if this would affect the calculation of any other parameters like recombination rate rho or theta.

Thanks!

[benjeffery]

There's more detail about ploidy here if you missed it: https://tskit.dev/msprime/docs/stable/ancestry.html#sec-ancestry-ploidy
In addition to storing the individuals correctly in the tree sequence this argument also affects the time scale on which coalescence occurs.
If you have further questions please ask, I'm converting this issue to a discussion.

[isinaltinkaya]

Thank you very much!

I actually wanted to understand the expected behavior compared to ms/msms. I have been using msms before I moved on to using msprime. I have an equivalent demographic model defined for both msms and msprime, and with msms I was using nsamples*2 to simulate diploid samples. Would this be equivalent to using nsamples*2 in msprime or ploidy=2?

[isinaltinkaya]

I am a bit confused here. When I use [20,20,20], with write_ms I get 60 rows in binary haplotypes part of ms output. If I use [10,10,10] this will be 30, but I would like to get 2 haplotypes per individual to represent diploid individuals.

[jeromekelleher]

Can you paste in your code please @isinaltinkaya, I think I'm probably misunderstanding things here.

[isinaltinkaya]

I use msprime to simulate samples=[20,20,20] and ploidy=2. Then I convert ts file ms format using write_ms.

When I check the number of haplotypes in the ms output, there are 60 haplotypes:

$ tail -n+7 test.ms |wc -l
60

But since my individuals are diploids, I expected to get 120 lines here (2 lines per individual).

[jeromekelleher]

Can you put in the full Python code called to sim_ancestry and write_ms please?

[isinaltinkaya]

I just found out that I had the same confusion as described in #1761; therefore it is resolved. Thanks a lot!