PathSeq freezing

Follow

John Perkins

• November 02, 2023 18:08
• Edited

Hello all, I am a new GATK user. I'm having problems getting the PathSeq pipeline to run, either by using PathSeqPipelineSpark or by running each of the individual steps. There are no error messages generated, it just freezes mid run. I am using an HPC blade w/ plenty of memory, added a java options flag to up the max memory, I've tried running w/ smaller portions of my dataset and databases, and I've validated my .BAM files w/ picard ValidateSamFile. Running through each step, it seems to be getting stuck on PathSeqBwaSpark (step 2). I should also note that I did get PathSeqPipelineSpark to work with the tutorial data, and it seemed to function without issue. For reference, my GATK version is 4.2.4.0.

 

0

• 

  Mark Walker

  • November 17, 2023 17:02

  Hi John Perkins,

  BWA will use off-heap memory to load the microbe index, so you want to make sure you have plenty of headroom for that. I'd suggest reducing the Java heap memory to the total memory less the size of the index, plus a little extra for the OS (a few GB). You may be running into garbage collection problems causing it to freeze.

  0

  Comment actions Permalink

Please sign in to leave a comment.