Post date: Mar 21, 2015 5:27:0 PM
I have completely reworked popanc to use LDA to calculate Pr(G | Z), estimate local ancestry per individuals, and incorporate uncertainty in the scale parameter. I am going to try the new version on new simulated data sets that I think will better capture relevant parameter space for the model. Here are the new conditions:
Number of generations = 20, 50, 200
Population size = 500
Number of markers per chromosome = 10001
Number of chromosomes = 10 simulated, but I will analyze two to keep the analyses more reasonable
Sample size = 50 per population (admixed and parental)
The simulations will be here: /labs/evolution/projects/popanc_sims/sims/.
I still need to think of how I want to then sample allele frequencies. I want to avoide low MAF markers.