Table 2.

Summary of challenge top performers

Technology	Genomic region	Participant	Performance metrics			F1 Rank
			F1	Recall	Precision	All	Diff	MHC
MULTI	alla	Sentieon	0.999	0.999	0.999	1	4	1
MULTI	alla	Roche Sequencing Solutions	0.999	0.999	0.999	1	1	7
MULTI	alla	The Genomics Team in Google Health	0.999	0.999	0.999	1	2	4
MULTI	diff	Roche Sequencing Solutions	0.994	0.992	0.996	1	1	7
MULTI	MHC	Sentieon	0.998	0.998	0.998	1	4	1
ILLUMINA	all	DRAGEN	0.997	0.996	0.998	1	1	5
ILLUMINA	diff	DRAGEN	0.969	0.961	0.978	1	1	5
ILLUMINA	MHC	Seven Bridges Genomics	0.992	0.989	0.996	6	9	1
PACBIO	all	The Genomics Team in Google Health	0.998	0.998	0.998	1	2	4
PACBIO	diff	Sentieon	0.993	0.991	0.994	4	1	1
PACBIO	MHC	Sentieon	0.995	0.993	0.997	4	1	1
ONT	all	The UCSC CGL and Google Health	0.965	0.947	0.984	1	1	2
ONT	diff	The UCSC CGL and Google Health	0.983	0.976	0.988	1	1	2
ONT	MHC	Wang Genomics Lab	0.972	0.964	0.980	3	3	1

One winner was selected for each technology/genomic region combination, and multiple winners were awarded in the case of ties. Winners were selected based on submission’s F1 score for the semi-blinded samples, HG003 and HG004 (harmonic mean of the parents’ F1 scores for combined SNVs and INDELs). Overall submission rank for all three genomic categories indicates submission overall performance: all, all benchmark regions; diff, difficult-to-map regions.

^aTie.