agshumate/Liftoff

Excessive memory usage with the polish option

Opened this issue · 0 comments

Thanks for the excellent tool! When running Liftoff between tomato genomes (~800 Mbp, 35k genes) with the following command:

liftoff Slycopersicum_691_SL4.0.fa Slycopersicum_390_v2.5.fa -g Slycopersicum_390_ITAG2.4.gene_exons.gff3 -o lifted.gff3 -p 32 -polish

it starts consuming a lot of memory and eventually crashes when the memory usage exceeds 100 Gb. Stderr log is here: liftoff_stderr.txt

To reproduce, you can download the input files from JGI's Data Portal at https://data.jgi.doe.gov/refine-download/phytozome?organism=Slycopersicum&expanded=Phytozome-390%2CPhytozome-691. I'm getting the error with the latest Liftoff (v1.6.3) and minimap2 (v2.24-r1122) releases.

Everything works fine without the -polish option, so this is not a pressing issue. Just wanted to make a record of this.

Best,
Tomas