strange result about the estimated cross-coalescence rate

[liuqi-C]

trafficstars

When I use MSMC2 (msmc_v2.1.4) to calculate the cross-coalescence rate with different time segment setting, I get some strange result, shown as follow.

1. --quantileBoundaries -p 10*1+15*2
2. -p 1*4+25*2+1*4+1*6
3. default setting. This result seems like the correct result, but the time interval is too little before 10000 years ago, only 4 time intervals.

How can I get a correct result?

[stschiff]

Well, the first analysis just has some bizarre blow-up effect in a time-window, which doesn't look good. If you plot it on a scale from 0 to 1 you should probably get something that looks a bit like the other two plots.

I can't really advise what to do here. It depends on a lot of things, possibly simulations would be necessary to see how its inference performs in various scenarios. Hard for me to say anything more specific, I am afraid.