I have a weird problem with MSMC2. I use it to study demography of a non-model species, with populations diverged over the last 100,000-1M years (the times are deeper than for the usual human work).
In almost all cases, MSMC based on SHAPEIT-phased haplotypes recovers ridiculous increases in Ne close to the present, like an order of magnitude boom in the last 10Kyr (last 4-5 slice of time, corresponding to the first part of the SMC command). This seems very unlikely and does not correlate with differences between the populations (plus the inferred Ne values are way to high for this animal). The middle of the curves looks reasonable, and consistent with PSMC on unphased data. The only solution I have had for this is to tie the five most recent slices together, so the Ne stays flat.: -p 1 4+25 1+1 2+1 5 [markodwn: there are normally asterisks instead of the spaces.]
Similar weird thing happens in the deep past - the inferred Ne for the first time slice is ridiculously high. I don't even understand how that happens with the coalescent.
What could cause these behaviours? I don't care about the very recent or the very deep past, but why is this happening?...