BUSCO v4 Running Time longer compared to v3
I was wondering if the following running time is normal for the new BUSCO v4, and if there's anyway to improve the speed.
BUSCO version: 4.0.2
blast version: 2.2.31
Number of cores: 20
RAM available: 220G
Genome evaluated: GRCh38.p13
mammalia_odb10: 9226 BUSCO groups -> 5.6 days
primate_odb10: 13780 BUSCO groups -> >6 days
Since BUSCO reads and writes a lot of small files, the database files and output folder are all on the local drive of the compute node instead of being accessed through the network drive. I noticed even though the blast version < 2.4, it still only uses 1 thread for the majority of the time. In comparison, running BUSCO v3 with mammalia_odb9 (4104 BUSCO groups) on GRCh38.p13 finished ~9hrs.
As we plan to run BUSCO on a number of human and mammalian genomes, any reduction in run time would be greatly appreciated.