FRED-2 / OptiType

Precision HLA typing from next-generation sequencing data

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

OptiType mapping memory issues

geejaytee opened this issue · comments

I'm running OptiType through a docker container (set up with 16GB memory on a 32GB Mac), and the initial mapping steps on my paired-end FASTQ data is failing, I suspect due to memory:

$ docker run -v <etc>
mapping with 1 threads...
 0:00:00.34 Mapping Sample_R1.fastq.gz to GEN reference...
Killed
 0:01:22.12 Mapping Sample_R2.fastq.gz to GEN reference...

The (GZ-)FASTQs are 150GB in size (I'm not using the mapped BAM files as they are >500GB each).

Is there a memory requirement to run the first step?

We're seeing this too, we ran through 60 samples of T/N data, half of which is failing for this step when setting the container memory to 16Gb. I'll see if I can mine the data on this and build a correlation plot of input file sizes and the expected memory usage for this step.

We also first convert our aligned bams back to fastq and realign specifically to the HLA region through razers3 (and then back to fastq again) before putting them through the optitype step.