Table 2.

The software selected for the performance test of genome annotation

Applicable object	Softwarea	Software release type	Average running speedb	Source
Prokaryotic genome	PROKKA	Stand alone	About 2 min	https://github.com/tseemann/prokka
	RASTtk	Stand alone	About 4 min	https://github.com/TheSEED/RASTtk-Distribution/releases/
	RAST	Web based	About 40 min	https://rast.nmpdr.org/
	GeneSAS_genemarkS	Web based	Less than 1 h	https://www.gensas.org/
	PGAP	Stand alone	About 5 h	https://github.com/ncbi/pgap
Eukaryotic genome	Companion_web	Web based	About 10 h	http://companion.sanger.ac.uk
	Companion_cl	Stand alone	About 15 h	https://github.com/sanger-pathogens/companion
	GeneSAS_genemarkES	Web based	Less than 1 h	https://www.gensas.org/
	GAL	Stand alone	About 1 day	https://hub.docker.com/u/cglabiicb/
	GAAP	Stand alone	More than 1 day	http://gaap.hallym.ac.kr/
Viral genome	VADR	Stand alone	About 1 min	https://github.com/nawrockie/vadr
	VAPiD	Stand alone	About 1 min	https://github.com/rcs333/VAPiD
	GeneSAS_genemarkS	Web based	A few minutes	https://www.gensas.org/
	GeneSAS_glimmer3	Web based	A few minutes	https://www.gensas.org/

^aFor stand-alone software except for PGAP, they are tested at a workstation with 8 CPU cores and 15 GiB memory. Because high-performance computing (at least over 4 GB memory/CPU) is required for PGAP, PGAP is tested at a server with 16 CPU cores, 376 GiB memory.

^bThe average of three running times of all related genome annotations.