burn-in before pedigree

[JaneeshBansal]

Hi,

I have a pedigree which I want to run for the final 4 generations of my simulation. But I would like to run a burn-in prior to the pedigree.

I have seen that the pedigree is usually specified as the initial state when simulating the ancestry. But I want to specify particular allele frequencies of the population before the pedigree. Would I need to extract the tree sequence from the pedigree simulation and then run recapitate?

Thank you!
Janeesh

[petrelharp]

I don't think there is a way to use msprime to simulate with given initial allele frequencies. The reason is that this is contrary to how msprime works fundamentally: msprime works by simulating backwards in time, from today towards the past; so specifying allele frequencies at some point in the past is not a natural thing to do at all.

More generally, I don't know of a good way to simulate genomes conditional on allele frequencies at a fixed set of positions (at least, simulate from a legit stochastic process, as opposed to just shuffling up nucleotides somehow).

[gregorgorjanc]

@JaneeshBansal I imagine you want to recreate what might have happened in reality - where someone created crosses (pedigree) with parents that had certain alleles or genotypes. I assume that these parents were picked out of a larger population, and then crossed to build the pedigree you now have. One way to simulate this would be to run msprime with many samples, to get the population. From this population select samples that match your criteria, (possibly simplify the tree sequence), then use SLiM in forward mode through pedigree (I assume this is possible - if not, you could pass haplotypes into AlphaSimR and drop them forward in time through your pedigree).

[gregorgorjanc]

I think you want to recapitate the pedigree tree sequence before dropping mutation.

[JaneeshBansal]

Can I recapitate on a msprime tree sequence? I tried doing it before adding the mutations on and I got this message TypeError: byte indices must be integers or slices, not str

I think the metadata needs to be in a SLiM style.

I might use msprime to initialise in SLiM and then run my pedigree.

Thank you for the suggestions!

[petrelharp]

Hi, @JaneeshBansal -

Can I recapitate on a msprime tree sequence? I tried doing it before adding the mutations on and I got this message: TypeError: byte indices must be integers or slices, not str

if you'd like input on why a particular error occurred or what to do about it, please provide the code you ran that produced the error and the full context of the error. Otherwise, we're just guessing.

[gregorgorjanc]

Can I recapitate on a msprime tree sequence? I tried doing it before adding the mutations on and I got this message TypeError: byte indices must be integers or slices, not str

@JaneeshBansal read these docs (they are examples in there):

• https://tskit.dev/msprime/docs/stable/pedigrees.html#basic-example
• https://tskit.dev/msprime/docs/stable/pedigrees.html#demography-information
• https://tskit.dev/msprime/docs/stable/ancestry.html#fixed-pedigree
• https://tskit.dev/msprime/docs/stable/ancestry.html#sec-ancestry-models-fixed-pedigree-completing
• https://tskit.dev/msprime/docs/stable/ancestry.html#pedigrees-and-demography