Astronomical increases in 5th generation

[fdrive98]

I am working with non-human primates datasets (with around 30k snps per chromosome). The results for the first generations are good and as expected, but they massively increase by the fifth generation in 2 of the 4 populations. Is there a way to solve it? How?

[armando-caballero]

Hi,
The first four generations have always the same estimated value because the method works in blocks of generations. If you have a very low value of Ne in these, and a huge increase from the 5th onwards, this is likely to be an artefact because of admixture of populations (you may see this type of figure in the tutorial). You may carry out a PCA or other method to see the possible estructuring in your sample and, perhaps, remove some individuals if necessary, although this will not guarantee success, as the admixture may have ocurred in previous generations.
Best wishes,
Armando.