ExaML bootstrap script?

[Tulio de Oliveira]

Hi 

Please let me know if it is an easier way to do bootstrapping in the ML trees using ExaML MPI version?

I could find some documentation on it (ExaML.pdf) that mention creating 100 parsimonies starting trees, but I could not find any information on the command line for running them in ExaML.

Thanks and kind regards

Tulio

[Muriel Gros-Balthazard]

Hi !

You need to use RaxML for this part.
Just use Examl when you want to obtain the tree but for the rest it's RaxML.

Here is how I do (please if someone can tell if it is OK !)

1) Create input phylip.binary
Using parser in the program ExaML

2) Generate 100 bootstrap replicates (BS)
raxmlHPC -# 10 -b 12345 -f j -m GTRGAMMA -s $phy.phylip -n $i\.REPS

3) Compute starting trees for each BS
raxmlHPC-PTHREADS -T 2 -y -s ../2_BSrep/$phy.phylip.BS$r -m GTRCAT -n vc_$i\_BS$r -p 12$i$r

4) Run ExaML with mpi
mpirun -n 16 \
    examl \
    -s input.phylip.binary \
    -n examlOut \
    -m GAMMA \
    -t RAxML_parsimonyTree.vc_$SGE_TASK_ID \
    -S

5) Combine all trees in TREES file and calculate bootstrap
raxmlHPC \
    -f b \
    -m GTRGAMMA \
    -s input.phylip.binary \
    -z TREES \
    -t one_tree \
    -n Bootstrap_tree

Hope this helps !

Muriel

[Alexandros Stamatakis]

Thanks Muriel,

Tulio, see my comments below and let me know if you have further questions.

> You need to use RaxML for this part.
> Just use Examl when you want to obtain the tree but for the rest it's RaxML.
>
> Here is how I do (please if someone can tell if it is OK !)
>
> 1) Create input phylip.binary
> Using parser in the program ExaML

:-)

>
> 2) Generate 100 bootstrap replicates (BS)
> raxmlHPC -# 10 -b 12345 -f j -m GTRGAMMA -s $phy.phylip -n $i\.REPS

This will only generate 10 BS replicates, also it's better to use -N 100
instead of -# 100 to indicate the number of replicates since the
character # will be interpreted as commented line by some queue
submission systems.

> 3) Compute starting trees for each BS
> raxmlHPC-PTHREADS -T 2 -y -s ../2_BSrep/$phy.phylip.BS$r -m GTRCAT -n
> vc_$i\_BS$r -p 12$i$r

correct, only that you don't need to use the pthreads version here,
since the parsimony calculations are not parallelized with pthreads.

also, it's much better to use the AVX vesrion of RAxML which is much
faster (factor 3) for parsimony calculations.

> 4) Run ExaML with mpi
> mpirun -n 16 \
> examl \
> -s input.phylip.binary \
> -n examlOut \
> -m GAMMA \
> -t RAxML_parsimonyTree.vc_$SGE_TASK_ID \
> -S

Note that, -S turns on the memory saving option for gappy alignments, if
you don't have a gappy alignment, it should not be used.

> 5) Combine all trees in TREES file and calculate bootstrap

easy to do using the Linux cat command

> raxmlHPC \
> -f b \
> -m GTRGAMMA \
> -s input.phylip.binary \
> -z TREES \
> -t one_tree \
> -n Bootstrap_tree

here the BS support will be drawn onto the branches of the best scoring
ML tree, alternatively you can build consensus trees out of the BS
replicate trees using the corresponding RAxML options

alexis

>
> Hope this helps !
>
> Muriel
>
>
> On Friday, August 28, 2015 at 12:40:49 PM UTC+2, Tulio de Oliveira wrote:
>
> Hi
>
> Please let me know if it is an easier way to do bootstrapping in the
> ML trees using ExaML MPI version?
>
> I could find some documentation on it (ExaML.pdf) that mention
> creating 100 parsimonies starting trees, but I could not find any
> information on the command line for running them in ExaML.
>
> Thanks and kind regards
> Tulio
>

> --
> 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