Stuck at queries from RAXML run.

[Farhana Riaz]

Hi,

I am running RAXML PTHREADS version for rapid bootstrapping +ML search on my nucleotide dataset of around 1460 genes with alignment patters =17930 using my workstation (64 GB RAM+32 cores).My 100 bootstraps have now been completed in  a matter of three whole days (using 30 cores) and its now been 48+ hours where my terminal shows "starting ML search"  and my system monitori  also shows all cpus busy running the process.I used the following command to execute:

raxmlHPC-PTHREADS -f a -T 30 -p 12345 -m GTRCAT -x 12345 -s coregenes_alignedfasta.afa -n coreoutput -# 100 

*QUESTIONS*

1)I have not used -n T20  (which AFAIK is indicating 20 ML searches to be run)in the command. Instead i percieved it to be the name of output file.Is it problematic? How many ML searches will be run in this case?

2) Is the terminal supposed to show "Starting ML Searches" and no progress after that throughout? Do I wait or kill the process? The cpus show the process to be running without any error.

3) How can I run a separate ML analysis from the 100 bootstrapped trees output file obtained to get th best scoringML tree with the bootstrapped values displayed?

4) I wanted to run the analysis on a cluster at my university which has working hours from morning to 8pm only after which it resumes the next day.Is there a flag to pause and continue the search the other way? Kindly guide stepwise for this procedure.

5) For 18SrRNA sequences. If some sequences have AGCT whereas others have AGCU,would RAXML process it fine?

Regards,

Ms.Farhana

[Alexandros Stamatakis]

> I am running RAXML PTHREADS version for rapid bootstrapping +ML search
> on my nucleotide dataset of around 1460 genes with alignment patters
> =17930 using my workstation (64 GB RAM+32 cores).My 100 bootstraps have
> now been completed in  a matter of three whole days (using 30 cores) and
> its now been 48+ hours where my terminal shows "starting ML search"  and
> my system monitori  also shows all cpus busy running the process.I used
> the following command to execute:
>
> raxmlHPC-PTHREADS -f a -T 30 -p 12345 -m GTRCAT -x 12345 -s
> coregenes_alignedfasta.afa -n coreoutput -# 100

First please note that standards RAxML is not supported any more, you
might consider switching to RAxML-NG:

https://github.com/amkozlov/raxml-ng

> *QUESTIONS*
> 1)I have not used -n T20  (which AFAIK is indicating 20 ML searches to
> be run)in the command. Instead i percieved it to be the name of output
> file.Is it problematic? How many ML searches will be run in this case?

it will run 20 ML searches.

> 2) Is the terminal supposed to show "Starting ML Searches" and no
> progress after that throughout? Do I wait or kill the process? The cpus
> show the process to be running without any error.

That's the normal output, keep in mind that ML searches are much slower
than rapid bootstrap searches.

> 3) How can I run a separate ML analysis from the 100 bootstrapped trees
> output file obtained to get th best scoringML tree with the bootstrapped
> values displayed?

You can run ML searches separately and independently from the 100
bootstrapped trees (see manual for details).

> 4) I wanted to run the analysis on a cluster at my university which has
> working hours from morning to 8pm only after which it resumes the next
> day.Is there a flag to pause and continue the search the other way?
> Kindly guide stepwise for this procedure.

This is not possible with standard RAxML.

> 5) For 18SrRNA sequences. If some sequences have AGCT whereas others
> have AGCU,would RAXML process it fine?

Yes.

Alexis

>
> Regards,
> Ms.Farhana
>
> --
> 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/50826d1e-fad9-4484-87d8-4f12219141ccn%40googlegroups.com
> <https://groups.google.com/d/msgid/raxml/50826d1e-fad9-4484-87d8-4f12219141ccn%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

www.exelixis-lab.org

[Farhana Riaz]

Dear Sir,

Thankyou so much for taking out the time to reply to my queries. Such help is not usually provided on forums and leaves a layman totally stranded. This helped me alot.

Best Regards,

Farhana

To unsubscribe from this group and stop receiving emails from it, send an email to raxml+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/raxml/882143ab-a75e-9816-adaa-1ca1ce03d749%40gmail.com.

[Farhana Riaz]

Dear All,

Almost 3 weeks to my original query, my run is smooth at RAXML and this is the current status (attached screenshot). I was running rapid 100 bootstraps and 20 ML trees to get the best tree. 

1)What remaining steps are left for the run to complete and generate the final output files?

2) Roughly how long would it take if you compare dataset details and the time taken by starting steps?

Regards

[Farhana Riaz]

Screenshot attached here

[Alexandros Stamatakis]

Dear Fahrhana,

There is one final more thorough optimization left of the best found ML
tree as far as I remember, I'd assume that the analysis should be done
by now.

Alexis

On 03.01.21 20:49, Farhana Riaz wrote:
> Dear All,
> Almost 3 weeks to my original query, my run is smooth at RAXML and this
> is the current status (attached screenshot). I was running rapid 100
> bootstraps and 20 ML trees to get the best tree.
>
> 1)What remaining steps are left for the run to complete and generate the
> final output files?
>
> 2) Roughly how long would it take if you compare dataset details and the
> time taken by starting steps?
>
> Regards
>
> On Thu, Dec 10, 2020, 12:42 PM Farhana Riaz <farhana...@gmail.com
> <mailto:farhana...@gmail.com>> wrote:
>
> Dear Sir,
> Thankyou so much for taking out the time to reply to my queries.
> Such help is not usually provided on forums and leaves a layman
> totally stranded. This helped me alot.
>
>
> Best Regards,
> Farhana
>
> On Thu, Dec 10, 2020, 9:41 AM Alexandros Stamatakis
> <alexandros...@gmail.com

> <mailto:raxml%2Bunsu...@googlegroups.com>
> > <mailto:raxml+un...@googlegroups.com
> <mailto:raxml%2Bunsu...@googlegroups.com>>.

> > To view this discussion on the web visit
> >
> https://groups.google.com/d/msgid/raxml/50826d1e-fad9-4484-87d8-4f12219141ccn%40googlegroups.com
>
> >
> <https://groups.google.com/d/msgid/raxml/50826d1e-fad9-4484-87d8-4f12219141ccn%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
>

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

>
> --
> 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%2Bunsu...@googlegroups.com>.

> To view this discussion on the web visit

> https://groups.google.com/d/msgid/raxml/882143ab-a75e-9816-adaa-1ca1ce03d749%40gmail.com.

>
> --
> 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/CAKTvUf_v9iZZxo7fd3zqr1eb6N%3DkPqyn_-j%3D5E9F4V88H-0jgg%40mail.gmail.com
> <https://groups.google.com/d/msgid/raxml/CAKTvUf_v9iZZxo7fd3zqr1eb6N%3DkPqyn_-j%3D5E9F4V88H-0jgg%40mail.gmail.com?utm_medium=email&utm_source=footer>.