search

bayesiancook / pbmpi

phylobayes mpi
GNU General Public License v2.0
23 stars 9 forks source link

bpcomp program encounters a problem #32

Closed yingqiouc closed 9 months ago

yingqiouc commented 1 year ago

Hi, I performed the BI analysis using PhyloBayes MPI version 1.8c. The number of taxa was 31 and the number of sites was 137483. But, when I carried out the bpcomp program to check the convergence of two chains, no matter how many trees were generated, the maxdiff was always 0. The model that I used was CAT-GTR+dgam 4. In addition, I discarded the first 10% cycles as burn-in. Could someone please tell me what might cause this result? Thanks.

bayesiancook commented 1 year ago

Hi, first: can you give me the exact code version that you are using? To do this, you can go in the pbmpi directory and type the following command:

git log

this will give you the list of commits starting from the most recent one. I would like to have the number, date and text of the first commit that appears upon typing this command.

other question: how many independent chains did you run? and can you give me the exact command that you are using for bpcomp?

best, Nicolas L.

bayesiancook commented 9 months ago

closed