Post date: Jul 01, 2015 11:0:38 PM
I am also re-running the Uygur data. I took the opportunity to also solve the singular matrix problem. You can check if a matrix is singular by calculating its determinant, which would be 0. When this happens I add a minor jitter (random noise) to the matrix so that it is no longer singular. A warning message is given too. Here is one of the new runs:
cd /local/scratch/
sleep 15
popanc -o estpanc_uygur15snpChrom9Chain3.hdf5 -m 30000 -b 10000 -t 5 -n 15 -d 10 -s 1 -u 20 -l 0.05 -a 0.1 -z 1 /labs/evolution/projects/popanc_sims/humanUygur/infiles/genHan9.txt /labs/evolution/projects/popanc_sims/humanUygur/infiles/genFrench9.txt /labs/evolution/projects/popanc_sims/humanUygur/infiles/genUygur9.txt
scp estpanc_uygur15snpChrom9Chain3.hdf5 /labs/evolution/projects/popanc_sims/mcmc/