Generating consensus trees directly prior to run completion

3 views
Skip to first unread message

Ziv Lieberman

unread,
Aug 5, 2024, 1:23:09 AM8/5/24
to ExaBayes
Hi all,
I submitted an ExaBayes run to Cipres recently, which terminated due to running out of time at ~800k generations, though configured originally to finish at 1.5E6 generations. After examining the combined trace from the output at this point, the sampling seems sufficient (well-mixed, converged, high ESS) for all parameters.

I'd now like to generate consensus trees, but Cipres only has the option to do this after the run is entirely complete. While I could restart it and run it to completion, this seems like a waste of computing time if the chains are already good enough.

Is there a way to take the ExaBayes_topologies.run-n files and make consensus trees with ExaBayes directly, rather than as part of an entire configured analysis? (This would have to be done locally, I think, until Cipres can expand its options).
I have used consense from Phylip, but my understanding is that this is different from that EB does (e.g., I had to manually remove the Nexus formatting and concatenate the Newick strings in the EB output files into a monolithic input for Phylip consense, which loses the taxon names from the nexus 'translate' block, and there are no support values (proportion of trees with a node) - which I suspect EB is supposed to retain?)

My apologies if this is addressed in the manual and I just missed it or misunderstood! I can also request Cipres to add this option if it can be done through EB commands, once I understand what to request ;)

Thanks!
-Ziv

Alexandros Stamatakis

unread,
Aug 5, 2024, 3:47:32 AM8/5/24
to exab...@googlegroups.com, Wayne Pfeiffer
Dear Ziv,

You do have access to the ExaBayes log and tree files until the point
where it terminated, correct?

In that case you should be able to use any post-processing tool for
calculating your summary stats.

The pre- and post-processing utilities you can use with ExaBayes are
listed here:

https://cme.h-its.org/exelixis/web/software/exabayes/manual/manual.html#prepost

Hope this helps,

Alexis

PS: I am also ccing Wayne Pfeiffer from SDSC here.

On 05.08.24 06:06, 'Ziv Lieberman' via ExaBayes wrote:
> Hi all,
> I submitted an ExaBayes run to Cipres recently, which terminated due to
> running out of time at ~800k generations, though configured originally
> to finish at 1.5E6 generations. After examining the combined trace from
> the output at this point, the sampling seems sufficient (well-mixed,
> converged, high ESS) for all parameters.
>
> I'd now like to generate consensus trees, but Cipres only has the option
> to do this after the run is entirely complete. While I could restart it
> and run it to completion, this seems like a waste of computing time if
> the chains are already good enough.
>
> Is there a way to take the ExaBayes_topologies.run-/n/ files and make
> consensus trees with ExaBayes directly, rather than as part of an entire
> configured analysis? (This would have to be done locally, I think, until
> Cipres can expand its options).
> I have used /consense/ from Phylip, but my understanding is that this is
> different from that EB does (e.g., I had to manually remove the Nexus
> formatting and concatenate the Newick strings in the EB output files
> into a monolithic input for Phylip /consense/, which loses the taxon
> names from the nexus 'translate' block, and there are no support values
> (proportion of trees with a node) - which I suspect EB is supposed to
> retain?)
>
> My apologies if this is addressed in the manual and I just missed it or
> misunderstood! I can also request Cipres to add this option if it can be
> done through EB commands, once I understand what to request ;)
>
> Thanks!
> -Ziv
>
> --
> You received this message because you are subscribed to the Google
> Groups "ExaBayes" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to exabayes+u...@googlegroups.com
> <mailto:exabayes+u...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/exabayes/7d3f9bed-4842-4c8d-955c-29e191ae8710n%40googlegroups.com <https://groups.google.com/d/msgid/exabayes/7d3f9bed-4842-4c8d-955c-29e191ae8710n%40googlegroups.com?utm_medium=email&utm_source=footer>.

--
Alexandros (Alexis) Stamatakis

ERA Chair, Institute of Computer Science, Foundation for Research and
Technology - Hellas
Research Group Leader, Heidelberg Institute for Theoretical Studies
Full Professor, Dept. of Informatics, Karlsruhe Institute of Technology

www.biocomp.gr (Crete lab)
www.exelixis-lab.org (Heidelberg lab)

Ziv Lieberman

unread,
Aug 6, 2024, 3:22:24 AM8/6/24
to exab...@googlegroups.com
Hi Alexander,
Of course, thank you! I had just read this incorrectly. Thank you for pointing me back in the right direction!
-Ziv

You received this message because you are subscribed to a topic in the Google Groups "ExaBayes" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/exabayes/yRBlks5UydQ/unsubscribe.
To unsubscribe from this group and all its topics, send an email to exabayes+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/exabayes/84be4b69-daee-44db-9647-852f92baf8d3%40gmail.com.
Reply all
Reply to author
Forward
0 new messages