SNP Tree using asc-corr on CIPRES

[Brendan Pinto]

Hi all,

I am trying to run RAxML with the stamatakis ascertainment bias correction (asc-corr=stamatakis) on the CIPRES server (RAxML-HPC v.8 on XSEDE) and continue to repetitively get this error.

"

This is RAxML Worker Pthread Number: 3
You specified that you want to use a stamatakis or felsenstein ascertainment bias correction for partition 0 with name p1.
but did not specify a correction file for this partition in the partition file!

" 

I know this is not true, but I don't know how to fix it. I only have one partition for my unlinked SNPs and a corresponding correction file. I am doing something wrong, but frustration is getting the better of me.

Please help!

Thanks in advance!

-bjp

[Mark Miller]

Hi Brendan,

I would do two things.

First, you can try to post this in the raxml google group, along with your partition file which will almost surely produce a correct response.

Second, send me the file _jobinfo.txt for a representative failure that has been submitted in the last two weeks.

Mark 

[Brendan Pinto]

Hey Mark, 

Thanks for your quick response. I have also bugged the RAxML group about this issue!

Attached is the _jobinfo.txt document.

Best,

bjp

[Mark Miller]

Hi Brendan,

Thanks for that.

You need to call the asc correction files from the partition file.
You cna find an example of how to do it in this post:
https://groups.google.com/forum/#!topic/raxml/kF8Ou3bJYY4

Let me know if that doesn't help.
Best,
Mark

[Brendan Pinto]

Hey Mark, 

I think it was a step in the right direction. There is no longer any errors, but it still terminates the run. 

It just doesn't say why...

Attached some output files..

Sorry and thanks again,

bjp

[Mark Miller]

Hi Brendan,

There is a message in stderr.txt that suggests a problem with the partition.

When you want an ASC bias correction in a partition, the name of each file will be controlled by CIPRES,

and you have to be careful that your partition file calls that file name.

For the 1 file you used, the name will be asc~p1.txt, and is you use more than one, they will increment,

to asc~p2.txt, asc~p3.txt etc, as specified in the parameter interface for this tool.

So in your partition file, you must have this line:

[asc~p1.txt],ASC_DNA,p0=1-11452

not

[p2.txt],ASC_DNA,p0=1-11452

because you need the correct name for your asc bias file.

Let me know if this helps or not.

Mark

[Brendan Pinto]

Yeah sorry I keep changing things incrementally and running them and they can get a bit distorted over time.. I tried it few times..

I really don't know what this is implying? I am certain there aren't any..

custom protein substitution or ascertainment bias file [p1.txt] you want to use does not exist!
you need to specify the full path
the file name shall not contain blanks!

Again sorry for the hassle..

[Mark Miller]

Hi Brendan,

Thanks for making that test. It seems there was a bug in the raxml interface.

I believe I have fixed the bug now, but I cannot release it in the production site right away because my colleague who manages releases is away for a week.

However, you can access this fix on our test server if you need quicker access.

It is at http://billiards3.sdsc.edu/portal2

You will have to create an account and then bring in your data.

Let me know if you have any problems.

Best,

Mark

[Brendan Pinto]

IT WORKED! Almost crying I'm so happy...

One last issue is that there is no bipartition file associate with the output similar to the previous post on this forum.

Very exciting!! 

Thanks so much for all of your help!

[Mark Miller]

Hi Brendan,

I think the absence of bipartitions is because the search did not converge.

Below is the last line of the stdout.txt.

It could be you want to run more, or maybe the signal isn't strong enough in your data set?

Mark

Rapid BS search did not converge after 1008 replicates with MRE-based Bootstopping criterion

WRF Average of 100 random splits: 3.327922

Overall Time for 1008 Rapid Bootstraps 141.125227 seconds

Average Time per Rapid Bootstrap 0.140005 seconds

All 1008 bootstrapped trees written to: /projects/ps-ngbt/backend/comet_test_workspace/NGBW-JOB-RAXMLHPC8_XSEDE-FE22128513ED4401B16AD55C654923CB/RAxML_bootstrap.result

[Brendan Pinto]

Hey Mark,

So I think that I figured out what was going on with the dataset that I was having trouble with most likely the signal was not strong enough, so while I am working on that issue.. In the meantime, I proceeded to try to replicate my results with another dataset that does have strong signal and am proceeding to run into the first problems I was having before the adjustment you made with the test server, but the test server couldn't complete the job either...

Could you take a glance and let me know?

Thanks,

This is RAxML MPI Process Number: 0
Error external partition file type acs does not exist
Available file types: asc and prot
Error external partition file type acs does not exist
Available file types: asc and prot
Error external partition file type acs does not exist

[Mark Miller]

Hi Brendan, thanks for reporting that.

The jobinfo.txt file you sent seems to me to be from the production server.

It is true I forgot to release the fix into production. That should happen in the next couple of hours.

But if you are also having trouble with the test server, can you send me a jobinfo,txt file from the test server, an I will investigate.

The first file will be named p1.txt, and so your partition file must call for a file with that name.

Let me know if I can help further.

Mark

 

[Brendan Pinto]

Here is the _jobinfo.txt from the production server. Same input, same output.

Thanks Mark!

[Mark Miller]

I think the problem is in your part file. it has this

[acs~p1.txt],ASC_DNA,p1=1-7645

But it should be

[asc~p1.txt],ASC_DNA,p1=1-7645

Hence the error message:

type acs does not exist: Available file types: asc and prot

If you fix the typo it should work fine.

HTH

Best,

Mark

[Mark Miller]

BTW, the fix is now implemented in the production portal.

So you can run equally well in either place.

Best,

Mark