Using too much RAM (memory)

TreesLab / NCLscan

We have developed a new pipeline, NCLscan, which is rather advantageous in the identification of both intragenic and intergenic "non-co-linear" (NCL) transcripts (fusion, trans-splicing, and circular RNA) from paired-end RNA-seq data.

MIT License

6 stars 9 forks source link

Using too much RAM (memory) #30

Open kopardev opened 1 year ago

kopardev commented 1 year ago

Running NCLscan on our HPC runs into memory issues. The dashboard plots look like this:

Any idea what is happening and how to get around this issue? Host is hg38 + a few viruses (well annotated).

chiangtw commented 1 year ago

Hi,

It may be caused by the mp_blat.py, I guess.

mp_blat.py is just a wrapper for applying multiple blat processes, not a real parallelized version of blat, so the maximum memory usage would be equal to *4GB mp_blat_process**. ("4GB" is for human genome, and the mp_blat_process is a parameter in the NCLscan.config, which is for tuning the number of proceses for mp_blat.py)