Tracing recombination events

[jikhashkya]

What's the best way to track the time of the recombination events that gave rise to IBD segments' boundaries ? I followed this link for some insight and it was very helpful but I'm not entirely sure if the recombination events pertaining to all IBD segments' boundaries can be tracked successfully ? Any help would be much appreciated.

[jeromekelleher]

I think the best option is to use the record_full_arg option. This will stored nodes associated with each RE event, so you can trace back the source of each breakpoint to a particular event. I can help a bit more if you're having trouble pulling out this information.

[jikhashkya]

@jeromekelleher - Thank you for the help. As for the second part of my question, let's take for example sample pair 0 and 4 and the IBD segments shared by these two samples specified by the genomic position intervals (left , right) are:

1. (0, 3)
2. (3, 5)
3. (5, 6)
4. (6, 10)

Now, trying to relate the RE event that led to these IBD segments, would it be correct to say that (0,3) IBD segment was the result of RE: 22/23, (3,5) was the result of RE:22/23 on the left side and RE:15/16 on the right and similarly, for the rest of the IBD segments for this sample pair ? Additionally, for sample pair 0 and 3, one IBD segment covers the positions (0,5) (with MRCA 27) , so is RE:15/16 also responsible for this IBD segment ?

[jeromekelleher]

This isn't obvious to me I'm afraid @jikhashkya - is there always going to be a one-to-one correspondence between IBD segments and recombination events? I guess we can trace back the source of a given breakpoint to an event (or events), and you could say that these "cause" the specific IBD segments?

[jikhashkya]

I apologize if the question is getting away from msprime-specific discussions, but from my understanding the boundaries of an IBD segment are results of some ancient recombination events, but tracing back to those events might not be as trivial ?

[jeromekelleher]

I think it's tricky, because we have breakpoints in non-ancestral material they may also give rise to IBD segment breaks which align with breakpoints that were actually caused by recombinations in ancestral material. It's not obvious to me which event cause the breakpoint then, or how you'd disambiguate this situation (or, whether it actually matters at all!). I'm sure there's a way to do it, but I'd have to spend a good bit of time on it to be confident that it's the right answer.

Sorry I can't be more helpful!

[jikhashkya]

I appreciate your help anyways. Thank you @jeromekelleher