Can one simulate diploid individuals?

[shaharMazia]

Hi, I tried to address this question in Biostars and didn't get an answer.
You are welcome to read the question (I prefer not to copy it here, but if it will be suggest, I will).

The current idea I have to deal with it, is to just join randomly genomes from the output to create diploid individuals. OF course to join genomes from the same sub populations.

I am afraid though, my entire simulation, which has recombination, is not simulated correctly then.
Did anyone experience with backward simulations for diploids? Is that even possible with coalescent theory?

Thanks,
Shahar

[benjeffery]

Hi @shaharMazia!

You're using the legacy simulate function. It is highly recommended to use the updated methods of sim_ancestry and sim_mutations.
There is a ploidy argument to sim_ancestry(). Have you seen the section on ploidy in the docs which explains the effect of this in msprime?
https://tskit.dev/msprime/docs/stable/ancestry.html#sec-ancestry-ploidy

[shaharMazia]

Hi @benjeffery thanks for the very rapid response!
I noticed the ploidy arg for sim_ancestry(), but I am using msprime.simulate().
I am not sure what is the different, I learned to use the lib through this paper that uses the simulate() method.
Can I use sim_ancestry() instead of simulate() and the rest will be the same? Sounds like a great solution if thats all, but some guts tells me backward simulation with recombination suppose to be a lot harder.

The docs says "The number of monoploid genomes per sample individual
(Default=2). See the :ref:sec_ancestry_ploidy section for usage examples."

It sounds like it is not exactly like diploid, but having 2 monoploid genomes per individuals (a bit like the hack I thought to do after the simulation). I couldn't find sec_ancestry_ploidy section, can you help me with that?

Thanks a lot, Shahar

[shaharMazia]

Oh I just saw you answered as well in the Biostar post.
Thanks, I will read a bit, and answer you there if I have more questions.

[benjeffery]

Yeah, I answered in both places in case someone else sees it over at BioStars. simulate was deprecated and replaced with sim_ancestry and sim_mutations as these fix some deficiencies and simplify the API. A paper updating the msprime one you linked to is at https://www.biorxiv.org/content/10.1101/2021.08.31.457499v3. msprime does the right thing with respect to polidy - it's not a hack!

The ploidy section of the docs is at this link: https://tskit.dev/msprime/docs/stable/ancestry.html#sec-ancestry-ploidy. msprime does the