unlink brlens=(all) for the MrBayes block

[Athulya Kizhakke]

Hello,

MrBayes is generating two ".con" trees (tree1.con.tre and tree2.con.tre) when I run the analysis using the MrBayes block from the PartitionFinder output. On googling why this could be, I happened to find a solution on Biostars which suggested that the "unlink brlens=(all)" option is what causes this error. Even though  it did not make sense to me after having gone through the description of the "unlink" command, when I re-run the analysis after deleting the line from the MrBayes input config file, only a single ".con" file was generated. Does anyone know what could be happening? Please find my MrBayes input block below (lines in red are from the PartitionFinder best_scheme file)

begin mrbayes;
    set autoclose=no nowarn=no;
    execute ConcatenatedAlignment.nex;
   
    charset Subset1 = 3546-3930\3 2553-3543\3 1-417 1873-2550\3 418-1803\3 1804- 1872 1874-2550\3 419-1803\3 3544-3930\3 2551-3543\3 3545-3930\3 2552-3543\3;
    charset Subset2 = 1875-2550\3 420-1803\3;

    partition PartitionFinder = 2:Subset1, Subset2;
    set partition=PartitionFinder;

    lset applyto=(1) nst=6 rates=invgamma;
    lset applyto=(2) nst=6 rates=invgamma;

    prset applyto=(all) ratepr=variable;
    unlink statefreq=(all) revmat=(all) shape=(all) pinvar=(all) tratio=(all);
    unlink brlens=(all);

    mcmcp Nruns=2 temp=0.1 Ngen=50000000 nchains=4 samplefr=1000 printfr=1000 diagnfr=10000;
    mcmc;
    sump burninfrac=0.25;
    sumt burninfrac=0.25;
end;

Thanks in advance!

Athulya

[Rob Lanfear]

Hi Athulya,

Linking or unlinking the branchlengths is a big difference in the two models you're implementing in MrBayes, and they may or may not end up being closely related to the models you estimated in PartitionFinder.

I'm guessing that when you unlink the branchlengths you get two trees which represent the two independent sets of branchlengths (one for each subset).

Rob