Closed pengyan19 closed 2 years ago
Hi @pengyan19,
It looks like the upper bound and starting parameters were missing from these models. The f parameter needs to be bounded between 0 and 1 for the model to make sense. I added in the appropriate bounds in the dadi_Run_3D_Set.py
script. If you download the updated version, you will need to re-run the admixture models using these updated settings. Apologies for any inconvenience, and thanks for posting this issue.
For transformations to true time and other parameters, please consult the dadi manual and the dadi user group.
For example, we used the admix_origin_no_mig model in 3 population, but i got the best AIC result which is [4.3764,4.0015,1.4507,0.0146,0.0117,1.3506],["nu1, nu2, nu3, T1, T2 , f"), why the f >0? and how did the T1,T2 transform to true time?