Variable results

[Anonymous]

Hi,

I have a question about the results I am getting when running BPP/3.2.
I am using 5 groups and trying to determine the species delimitation of them. I have a specified guide tree to use and I'm running the 10 analysis so my tree is fixed. However, my results for the "Guide tree with posterios probability for presence of nodes" are always extremely variable when I run multiple instances with the same .ctl file.

10_1-(((TA, (TS, TN) #0.029720) #0.225360, TU) #1.000000, TB) #1.000000;
10_2-(((TA, (TS, TN) #0.180170) #0.489880, TU) #1.000000, TB) #1.000000;
10_3-(((TA, (TS, TN)), TU), TB) #1.000000;
10_4-(((TA, (TS, TN)), TU), TB);
10_5-(((TA, (TS, TN) #0.413600) #1.000000, TU) #1.000000, TB) #1.000000;

I understand there should be some variety in my results, but this seems to be too much to have accurate results.

I tried to create an unreasonable tree, hoping to see a decrease in the probability of nodes, but I get close to the same thing. In addition, the outermost nodes are all at 100%.

10_1-((((TN, TB) #0.250400, TA) #0.570400, TU) #1.000000, TS) #1.000000;
10_2-((((TN, TB), TA), TU), TS) #1.000000;
10_3-((((TN, TB) #0.263610, TA) #0.555780, TU) #1.000000, TS) #1.000000;

My MCMC moves seem to be fine, along with my priors for theta and tou. Do you have any suggestions as to why I might be getting such different values?

Thanks.

[Rannala]

I suggest that you try running the program inferring the species tree as well (using option "speciestree = 1 1 0.5 0.1 0.1" in the control file) rather than using a guide tree.

[Anonymous]

I am having the same problem. Any thoughts about why this is happening, or how we could try to fix it?
Run 1: ((N, M)'#1.00000', S)'#1.00000'
Run 2: ((N, M)'#30619', S)'#0.30910'
Run 3: ((N, M)'#0.01306', S)'#1.00000'