Rapid bootstrapping completed, but no trees?

[Breanna Sipley]

Hi there,

It was my assumption that if no -f was specified RAxML would execute the standard hill climbing algorithm (-­f d). I have run RAxML before under this assumption (w/o including the -f option), & it ran as expected. However, more recently, RAxML appeared to quit after executing rapid bootstrapping & writing my 1000 bootstrapped trees.

I called RAxML as follows:

raxmlHPC-PTHREADS-AVX2 -T 20 -m PROTCATBLOSUM62 -n OUTPUTNAME -s ALIGNMENT -# 1000 -p 12345 -x 12345

The final lines in my *info* are:

Overall Time for 1000 Rapid Bootstraps 36326.376972 seconds
Average Time per Rapid Bootstrap 36.326377 seconds
All 1000 bootstrapped trees written to: RAxML_bootstrap.OUTPUTNAME

Did RAxML not finish because I didn't include -f d? I received no error messages. Can I execute a thorough tree search on those bootstrapped trees & generate the RAxML bipartition files? I tried including -f T -t RAxML_bootstrap.OUTPUTNAME, but it appears this only gives me a RAxML_bestTree.*

Cheers,
Breanna

[Alexey Kozlov]

Hi Breanna,

in general, you can use the "-f a" option to run this "all-in-one" analysis (ML tree search + rapid bootstrapping +
bootstrap bipartition support mapped to the best-found ML tree).

But you can also obtain the bipartition support from your existing output files, please see detailed explanations
between the lines:

> It was my assumption that if no -f was specified RAxML would execute the standard hill climbing algorithm (-­f d).

It is generally true, but there are few exceptions, including "-x" and "-b" options.

> I have run RAxML before under this assumption (w/o including the -f option), & it ran as expected. However, more recently,
> RAxML appeared to quit after executing rapid bootstrapping & writing my 1000 bootstrapped trees.
>
> I called RAxML as follows:
>
> |

> raxmlHPC-PTHREADS-AVX2 -T 20-m PROTCATBLOSUM62 -n OUTPUTNAME -s ALIGNMENT -# 1000 -p 12345 -x 12345

> |
>
> The final lines in my *info* are:
>
> |
> OverallTimefor1000RapidBootstraps36326.376972seconds
> AverageTimeper RapidBootstrap36.326377seconds

> All1000bootstrapped trees written to:RAxML_bootstrap.OUTPUTNAME

> |
>
> Did RAxML not finish because I didn't include -f d? I received no error messages.

RAxML did finish successfully and generated rapid bootstrap trees.

>Can I execute a thorough tree search
> on those bootstrapped trees & generate the RAxML bipartition files? I tried including -f T -t
> RAxML_bootstrap.OUTPUTNAME, but it appears this only gives me a RAxML_bestTree.*

This is also intended behavior. Now you just need to run one extra command to map BS support values to the best ML tree
("-f b" option):

|
raxmlHPC-AVX2 ­-f b ­-t RAxML_bestTree.OUTPUTNAME ­-z RAxML_bootstrap.OUTPUTNAME ­-m GTRGAMMA ­-n SUPPORT
|

Please also refer to the RAxML manual which describes what each particular command line option is doing exactly:

https://sco.h-its.org/exelixis/resource/download/NewManual.pdf

Hope it helps,
Alexey

[Breanna Sipley]

Hi Alexey,

This helped a ton! It's great to know how to find the best ML tree and map BS support onto it independently from generating rapid bootstrap trees.  It's fun when what you think is unintended behavior is actually a feature!

Thanks!

Cheers,

Breanna

--
You received this message because you are subscribed to a topic in the Google Groups "raxml" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/raxml/Ots0gZEZfx0/unsubscribe.
To unsubscribe from this group and all its topics, send an email to raxml+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Alexey Kozlov]

you're welcome :)

On 03.05.2018 17:13, Breanna Sipley wrote:
> Hi Alexey,
>
> This helped a ton! It's great to know how to find the best ML tree and map BS support onto it independently from
> generating rapid bootstrap trees.  It's fun when what you think is unintended behavior is actually a feature!
>
> Thanks!
>
> Cheers,
> Breanna
>

> <https://groups.google.com/d/topic/raxml/Ots0gZEZfx0/unsubscribe>.
> To unsubscribe from this group and all its topics, send an email to raxml+un...@googlegroups.com
> <mailto:raxml%2Bunsu...@googlegroups.com>.
> For more options, visit https://groups.google.com/d/optout <https://groups.google.com/d/optout>.
>
>
> --
> 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>.