Extremely small Ne
Closed this issue · 2 comments
Sometimes simulating IBD segment with simulate_ibd and its derivatives throw an error when constant Ne is less than 3000. This may be a small floating points issue. I will not look into this issue further unless prompted by a user who has an Ne(t) demography that causes such an issue.
This is a really important bug! It had to do with the function cut_Ne that is also called in the simulate_ibd_isweep function. I had the dictionary filtering based on the effective size (value), not the generational time (key).
I will have to recompile the package to fix this. I will close this issue once I have made the update.
This would not have affected any of my results. I always studied Ne's that were >= 2000 uniformly.
Fixed with new update to the package.