resuming rapid bootstrap run

[Francesco Cicconardi]

Hi everyone,

I'm currently running a rapid Bootstrap analysis and best-scoring ML tree (-f a) on my cluster, it is taking a lot of time and I'll probably exceed job's run time limit. Is there a way to resume the run?

thank yoy,

Francesco  

[Alexandros Stamatakis]

Unfortunately not,

But you can

1. just run a couple of rapid bootstraps per job (without search for the
best tree) and then cconcatenate the BS tree files

2. then also run a couple of standard ML searches independently and
check the RAxML_info files for the best scoring tree

3. then you can build consensus trees or draw BS support values on the
best-scoring tree with the respective RAxML options

Alternatively you can deploy the RAxML hybrid MPI-Pthreads version.

Finally, if the dataset is really huge you may consider using ExaML instead.

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>.
> For more options, visit https://groups.google.com/d/optout.

--
Alexandros (Alexis) Stamatakis

Research Group Leader, Heidelberg Institute for Theoretical Studies
Full Professor, Dept. of Informatics, Karlsruhe Institute of Technology
Adjunct Professor, Dept. of Ecology and Evolutionary Biology, University
of Arizona at Tucson

www.exelixis-lab.org

[Francesco Cicconardi]

Thanks Alexis,

So, let's see if options are right:

 

1. just run a couple of rapid bootstraps per job (without search for the

best tree) and then concatenate the BS tree files

option: "-f a"

 

2. then also run a couple of standard ML searches independently and
check the RAxML_info files for the best scoring tree

option: "-f d" ???

3. then you can build consensus trees or draw BS support values on the
best-scoring tree with the respective RAxML options

option: "-f b"

Alternatively you can deploy the RAxML hybrid MPI-Pthreads version.

Finally, if the dataset is really huge you may consider using ExaML instead.

I'll check this anyway! ;)

Thanks

F

[Alexandros Stamatakis]

On 22.01.2016 21:19, Francesco Cicconardi wrote:
> Thanks Alexis,
>
> So, let's see if options are right:
>
> 1. just run a couple of rapid bootstraps per job (without search for
> the
> best tree) and then concatenate the BS tree files
>
>
> option: "-f a"

yes but make sure to provide different random number seeds

> 2. then also run a couple of standard ML searches independently and
> check the RAxML_info files for the best scoring tree
>
>
> option: "-f d" ???

yes, you can also omit -f d since this is the default anyway

> 3. then you can build consensus trees or draw BS support values on the
> best-scoring tree with the respective RAxML options
>
>
> option: "-f b"

yes, that's for drawing BS support onto the best-scoring ML tree, if you
want to build a consensus out of the BS replicates use the -J option ...

alexis

>
> Alternatively you can deploy the RAxML hybrid MPI-Pthreads version.
>
> Finally, if the dataset is really huge you may consider using ExaML
> instead.
>
>
> I'll check this anyway! ;)
>
> Thanks
> F
>
>
> Alexis
>
> On 19.12.2015 18:57, Francesco Cicconardi wrote:
> > Hi everyone,
> >
> > I'm currently running a rapid Bootstrap analysis and best-scoring ML
> > tree (-f a) on my cluster, it is taking a lot of time and I'll
> probably
> > exceed job's run time limit. Is there a way to resume the run?
> >
> > thank yoy,
> > Francesco
> >
> > --
> > 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 <javascript:>
> > <mailto:raxml+un...@googlegroups.com <javascript:>>.

> > For more options, visit https://groups.google.com/d/optout

> <https://groups.google.com/d/optout>.

>
> --
> Alexandros (Alexis) Stamatakis
>
> Research Group Leader, Heidelberg Institute for Theoretical Studies
> Full Professor, Dept. of Informatics, Karlsruhe Institute of Technology
> Adjunct Professor, Dept. of Ecology and Evolutionary Biology,
> University
> of Arizona at Tucson
>

> www.exelixis-lab.org <http://www.exelixis-lab.org>

[Francesco Cicconardi]

Hi Alex,

Now I'm dealing a bigger dataset, I can follow your advice for bootstrap replicates, and it works, but what about the ML tree? Is there any feasible although less precise calculation of the tree? I'm computing three gene family trees and I don't really need super precise phylogeny at the moment, I'm more interested in the topology (the bootstrap will help) rather than the branch length.

Thank you

Francesco

[Alexandros Stamatakis]

could you tell me the dimensions of your dataset, i.e., #taxa and
#sites, are you using RAxML or ExaML?

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>.
> For more options, visit https://groups.google.com/d/optout.

--
Alexandros (Alexis) Stamatakis

Research Group Leader, Heidelberg Institute for Theoretical Studies
Full Professor, Dept. of Informatics, Karlsruhe Institute of Technology

www.exelixis-lab.org