possible to output two tree sequences?

[bob]

hi!

i'm curious if it's possible spit out two separate tree sequences from slim. for example, i might have two populations representing two different species. i would want to treat the two tree sequences as separate entities in downstream processing, including independent simplification and recapitation for each tree. has anyone tried this? any ideas?

thanks

bob

[Ben Haller]

SLiM can only output the tree sequence for the whole simulation, but you should be able to just simplify down to two different "samples" on the python side of things, I think?

Cheers,

-B.

Benjamin C. Haller

Messer Lab

Cornell University

[bob]

thanks for the response ben!

yea, that's what ive been doing (subsetting by "population", where "population" for me = species). however, this approach requires either the two species to have the same genetic architecture or for each species to carry a "dummy" chromosome which corresponds to the genome of the adjacent species. i haven't found a feature of tskit that allows for truncating chromosomes, just deleting information at intervals of loci or sites. it creates a lot of bookkeeping in downstream processing/analysis that would be nice to sidestep. perhaps this is more of tskit issue than a slim issue...

best

bob

[Ben Haller]

Well, it's a SLiM issue inasmuch as SLiM could let you simulate multiple species, with different genetics and so forth.  That's on the to-do list.  :->  But barring that rather large new feature, yes, the situation you're in is pretty much a tskit issue not a SLiM issue.  :->

Cheers,
-B.

Benjamin C. Haller
Messer Lab
Cornell University

bob wrote on 12/4/21 7:52 PM:

--
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/bce98c87-7ea1-4f88-bda9-b7c6934d933cn%40googlegroups.com.

[hyan...@googlemail.com]

Just to say that ts.rtrim() will truncate chromosomes from which the right hand end has been deleted using ts.delete_intervals() or ts.keep_intervals()

Yan

[bob]

thanks for the suggestion yan! i think this is exactly what i need :)

[Peter Ralph]

We deal with this by having different mutation types for each species,
and a script block that removes any mutations of the wrong type from
each species each generation. (However, they still have to have the
same genome length and recombination map, true - separate chromosomes
would be a nicer way to do it!)

> --
> 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/f00f77af-b01a-425a-9fad-f8be63da45b0n%40googlegroups.com.

[bob]

hey peter,

the way im doing it now incorporates yan's suggestion. i make two subpopulations, one for each species. i dedicate the first L1 loci (where L1 is a number) to species 1 and the next L2 loci to species two. i delete mutations landing on the "wrong" part of the chromosome (ie, the part of the chromosome corresponding to the species of which the focal individual is not a member) [im not sure if this step is necessary given my downstream processing]. in tskit, i subset the treequence by subpopulation (~species), use ts.delete_intervals() to delete information in the wrong parts of the chromosomes, and then ts.rtrim() for species 1 and ts.ltrim() for species 2 to isolate the chromosomes of each species. then i recapitate, mutate, output vcf, etc. i still need to triple check all the details, but i believe this results in cleanly separated chromosomes for each species. scripts available upon request.

best,

bob