Inconsistency in TBE values

[Berkay SELÇUK]

This is my initial command:

raxml-ng --threads auto{36} --tree pars{100} --workers auto --all --data-type AA --model JTT+F+R10 --site-weights /cta/users/bselcuk/GPCRA/ClassA/ClassA_tree/linsi_op4_ep0/7bins_exp_converted_range.txt --force msa_dups --force msa_names --seed 2 -msa /cta/users/bselcuk/GPCRA/ClassA/ClassA_tree/linsi_op4_ep0/consensus0.2_olfrep_wdup_linsi_op4_ep0_trimmedends_ginsi_op4_reordered_gt0.4.fasta --prefix ginsi_op4_gt0.4_7bins_exp_converted_range --bs-trees autoMRE{500} --bs-metric tbe,fbp

TBE measure is very important for my analysis. After getting poor TBE support values I wanted to validate it by using. 

raxml-ng --support --tree /cta/users/bselcuk/GPCRA/ClassA/ClassA_tree/ginsi_op4_gt0.4_7bins_exp.raxml.bestTree --bs-trees /cta/users/bselcuk/GPCRA/ClassA/ClassA_tree/ginsi_op4_gt0.4_7bins_exp.raxml.bootstraps --prefix 7bins_exp_validation --threads 1 --bs-metric tbe,fbp

When I compare FBP support values, they are the same on the other hand TBE are significantly different. I believe that they should be exactly the same too. 

Which one should I trust?

Can it be related to the rooting of the tree which I am doing when I am using Dendroscpoe? I am choosing the bootstrap option to visualize support values.

I would appreciate it if you could help. 

Thank you in advance.

[Alexey Kozlov]

Hello,

please send me all log and output files from both runs, and I will have a look.

Best,
Alexey

> --
> You received this message because you are subscribed to the Google Groups "raxml" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to
> raxml+un...@googlegroups.com <mailto:raxml+un...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/raxml/0864a700-0d64-4e99-bf48-bb4773a2aa8dn%40googlegroups.com
> <https://groups.google.com/d/msgid/raxml/0864a700-0d64-4e99-bf48-bb4773a2aa8dn%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Alexey Kozlov]

Hello again,

thanks for sending the files!

I can now confirm that there is indeed an inconsistency between both runs which is not supposed to
be there.

I'm still working on the fix, but for the time being, please use results from the validation run
(--support command) which which gives the correct values.

Another possible workaround is to use an alternative TBE algorithm "--extra tbe-naive" with the
--all command. It is somewhat slower, but on the dataset like yours the difference will be negligible.

I will keep you posted.

Best,
Alexey