Help Needed with PhyloBayes MPI Analysis on CIPRES

30 views

Skip to first unread message

Jee-Young

unread,

Nov 30, 2024, 9:04:15 PM11/30/24

to CIPRES Science Gateway Users

Hello,

I am currently using the "PhyloBayes MPI on ACCESS" tool on CIPRES to analyze mitochondrial PCG sequences with the CAT+GTR model.

Following the instructions provided in the "Advanced Help" section, I uploaded a .phy data file for my analysis and set the parameters as follows:

스크린샷 2024-11-30 152534.png

After completing the analysis, several output files were generated.

The bpcomp.con.tre file successfully produced a phylogenetic tree when downloaded.

However, upon checking the STDOUT file, I noticed that it only contained very brief content.

started chain 0 with PID : 2981333

started chain 1 with PID : 2981334

Terminator running as PID : 2981335

number of positions eliminated : 4896

Got SIGCHLD or SIGTERM, attempting graceful return.

This was surprising because, when using "MrBayes on ACCESS" for BI tree analysis, the STDOUT file was much more detailed. Is it normal for the STDOUT file in "PhyloBayes MPI on ACCESS" to be so short?

Additionally, the STDERR file contained the following messages:

Inactive Modules: 1) subversion/1.14.0

The following have been reloaded with a version change:

1) cpu/0.17.3b => cpu/0.15.4

model:

stick-breaking Dirichlet process mixture (cat)

exchangeabilities estimated from data (gtr)

discrete gamma distribution of rates across sites (4 categories)

read data from file : infile.phy

number of taxa : 23

number of sites : 8430

number of states: 4

chain name : chain0

run started

model:

stick-breaking Dirichlet process mixture (cat)

exchangeabilities estimated from data (gtr)

discrete gamma distribution of rates across sites (4 categories)

read data from file : infile.phy

number of taxa : 23

number of sites : 8430

number of states: 4

chain name : chain1

run started

srun: forcing job termination

srun: Job step aborted: Waiting up to 32 seconds for job step to finish.

slurmstepd: error: *** STEP 35460909.1 ON exp-3-05 CANCELLED AT 2024-11-29T20:47:54 ***

slurmstepd: error: *** STEP 35460909.2 ON exp-3-05 CANCELLED AT 2024-11-29T20:47:54 ***

srun: error: exp-3-05: tasks 0-31: Killed

/home/cipres/ngbw/contrib/tools/bin/PBMultiRun_expanse.bash: line 1: warning: run_pending_traps: bad value in trap_list[17]: 0x5555555c38c0

The message "forcing job termination" and several errors and warnings are mentioned, but I am unsure at which step the errors occurred.

Since this is my first time using "PhyloBayes MPI on ACCESS" for analysis, I would like to confirm the following:

Did I set the parameters correctly?
Are the generated results reliable?
If there are errors in the analysis and the results are incorrect, what might be the cause?
I would greatly appreciate any insights or advice from anyone familiar with this tool.

Thank you in advance for your help!

Best regards,
Jee-Young

Mark Miller

unread,

Dec 1, 2024, 12:28:55 PM12/1/24

to CIPRES Science Gateway Users

Hi Jee-Young,

Thanks for reporting the issue.

I do not see any jobs in CIPRES associated with this email address.

Can you send me the file _JOBINFO.TXT from this job, and I will investigate the job run.