Why raxml is taking is too long to generate raxml tree

145 views
Skip to first unread message

krishna acharya

unread,
Sep 8, 2022, 4:34:19 AM9/8/22
to raxml

Dear Alexis,
Good Morning!

I used raxmlHPC command for my alignment and it took me more than a month to align my 1.4 Mb genomic sequences of 779 isolates. So, I looked for a quicker way to generate raxml tree and I could see that raxmlHPC-PTHREADS-SSE3 is substantially faster. I submitted my alignment to the server with a total core of 16 cores and assigned 12 threads for my analysis. So, could you please kindly tell me how long it will take to get ML tree using this command? Will it be ready within a week?

raxmlHPC-PTHREADS-SSE3 -s Suberis_lcb_masked.fasta -m GTRGAMMA -f a -p 12345 -x 12345 -T 12 -N 1000 -n raxml_prasnp_concatenated_LCB_1055sites_masked_100boot.

Could you please let me know if there are any other alternatives to quickly prepare ML tree using raxml?

Thank you.



Krishna 

Alexandros Stamatakis

unread,
Sep 8, 2022, 11:37:24 AM9/8/22
to ra...@googlegroups.com
The execution time can not be predicted as RAxML optimizes the tree
iteratively and we never know how many iterations it will take.

What is helpful though is to first conduct just maybe one BS replicate
search and then extrapolate from there.

It is also helpful to know the alignment dimensions in order to
determine the approximate memory requirements and number of threads that
can be used.

Please read the extensive RAxML documentation at
https://cme.h-its.org/exelixis/web/software/raxml/index.html

and consider switching to RAxML-NG https://github.com/amkozlov/raxml-ng
as standard RAxML is not being supported any more.

Alexis
> --
> 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/253547d5-f7c4-43b1-a693-6af746053fban%40googlegroups.com
> <https://groups.google.com/d/msgid/raxml/253547d5-f7c4-43b1-a693-6af746053fban%40googlegroups.com?utm_medium=email&utm_source=footer>.

--
Alexandros (Alexis) Stamatakis

Research Group Leader, Heidelberg Institute for Theoretical Studies
Full Professor, Dept. of Informatics, Karlsruhe Institute of Technology
Affiliated Scientist, Evolutionary Genetics and Paleogenomics (EGP) lab,
Institute of Molecular Biology and Biotechnology, Foundation for
Research and Technology Hellas

www.exelixis-lab.org

Pfeiffer, Wayne

unread,
Sep 15, 2022, 10:50:21 AM9/15/22
to ra...@googlegroups.com, Pfeiffer, Wayne
Hi Krishna,

If you have not gotten your RAxML analysis to complete expeditiously, you could try the CIPRES gateway at www.phylo.org. This will schedule the analysis in a cost-effective way for the parameters of your data set, most likely on 128 cores. The first 1,000 core hours are free. After that, there is a modest cost for a subscription.

Bootstrap searches are much quicker with RAxML8 than RAxML-NG, so RAxML8 would be preferred if running time and cost are of concern. Also, it would be better to replace -N 1000 by -N autoMRE. This will do fewer than 1000 bootstraps if your data have a strong phylogenetic signal.

Good luck!

Wayne

On Sep 8, 2022, at 8:37 AM, Alexandros Stamatakis <alexandros...@gmail.com> wrote:

The execution time can not be predicted as RAxML optimizes the tree iteratively and we never know how many iterations it will take.

What is helpful though is to first conduct just maybe one BS replicate search and then extrapolate from there.

It is also helpful to know the alignment dimensions in order to determine the approximate memory requirements and number of threads that can be used.



Alexis

On 08.09.22 11:34, krishna acharya wrote:
Dear Alexis,
Good Morning!
I used raxmlHPC command for my alignment and it took me more than a month to align my 1.4 Mb genomic sequences of 779 isolates. So, I looked for a quicker way to generate raxml tree and I could see that raxmlHPC-PTHREADS-SSE3 is substantially faster. I submitted my alignment to the server with a total core of 16 cores and assigned 12 threads for my analysis. So, could you please kindly tell me how long it will take to get ML tree using this command? Will it be ready within a week?
raxmlHPC-PTHREADS-SSE3 -s Suberis_lcb_masked.fasta -m GTRGAMMA -f a -p 12345 -x 12345 -T 12 -N 1000 -n raxml_prasnp_concatenated_LCB_1055sites_masked_100boot.
Could you please let me know if there are any other alternatives to quickly prepare ML tree using raxml?
Thank you.
Krishna
-- 
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>.

-- 
Alexandros (Alexis) Stamatakis

Research Group Leader, Heidelberg Institute for Theoretical Studies
Full Professor, Dept. of Informatics, Karlsruhe Institute of Technology
Affiliated Scientist, Evolutionary Genetics and Paleogenomics (EGP) lab, Institute of Molecular Biology and Biotechnology, Foundation for Research and Technology Hellas


-- 
You received this message because you are subscribed to the Google Groups "raxml" group.
Reply all
Reply to author
Forward
0 new messages