search

FelixKrueger / Bismark

A tool to map bisulfite converted sequence reads and determine cytosine methylation states
http://felixkrueger.github.io/Bismark/
GNU General Public License v3.0
366 stars 101 forks source link

Bismark Extract failed always ran out of memory after 8hrs. #662

Open RachaelAjayi opened 3 months ago

RachaelAjayi commented 3 months ago

Bismark_mapping Failed, Run time 08:18:26, OUT_OF_MEMORY.

Please see code bismark --genome_folder genomes/bismark -o bam/ --non_directional --parallel 8 --score_min L,-0.6,-0.6 -1 trimmed/${base1}_val_1.fq.gz -2 trimmed/${base2}_val_2.fq.gz

What am I doing wrong? I am running on a cluster using 16 nodes for the job submission.

FelixKrueger commented 3 months ago

Assuming the human genome, and easy to work numbers, a single run of Bismark requires ~3 cores (at 100%) and ~14GB of RAM for a standard directional run, and 5 cores and ~20GB of RAM for a non directional run. Since you used --parallel 8 you need to roughly multiply these values with 2, meaning the run would be in the region of roghly 40 cores and up to 160GB. If you don't have that much, the job will be killed using due to a OUT_OF_MEMORY error.

Does that make sense?

RachaelAjayi commented 3 months ago

Thank you.

RachaelAjayi commented 3 months ago

Thank you so much for your reply.

New Error: 2- bowtie2-align died with signal 9 (KILL) (ERR): bowtie2-align died with signal 9 (KILL)

FelixKrueger commented 3 months ago

That still sounds like memory issues, what did you change? Can you google (ERR): bowtie2-align died with signal 9 (KILL)?