-
Notifications
You must be signed in to change notification settings - Fork 1
Description
Hello,
I am trying to calculate the log likelihood of a set of networks given a list of gene trees. The trees consist of 9 tips, all of which with no missing data. I collapsed all poorly supported branches (BS<50). I am trying to calculate log likelihood on networks with 0 to 5 reticulation events, and this analysis runs pretty quickly for networks with 0,1,2,3, and 5 reticulation events. However, I can never get the one with 4 to finish, even when using 64 CPUs, 4 GBs per CPU. My first run was with around 1950 gene trees, and then I decided that perhaps this was not computationally feasible, so I randomly selected 500, but my runs still go until they timeout. Is this truly not computationally feasible, or am I perhaps running into some bug that I cannot detect? I have attached these nexus files as text files below (NEXUS extension not supported). Thank you for your help.
500random_ret4_9_50col_getProb.txt
500random_ret3_9_50col_getProb.txt
Here is the run with 4 events with ~1950 gene trees: