Table 1. Selection of the optimal weight matrix by core genes and evaluation by the genome-wide association (GWA) p-values.

Parameter seta		Total number of weight matrices	Number of weight matrices met criteriab	Selection by core genes			Selection by GWA p-valuesf
ϕ (%)	η (%)			Weight matrixc	Position j d	Positionl e	Mean	sd
90	3	4,663	1	[2,1,1,8,1,1,7]	150	1053	966.6	5.4
	4	11,916	13	[2,1,1,8,1,1,7]	150	1053	963.9	4.1
				[2,1,1,8,1,1,8]	153	738	933.8	5.5
				[3,1,1,8,1,1,7]	153	1054	952.6	7.2
				[3,1,1,8,1,1,8]	155	738	930.2	9.4
				[4,1,1,8,1,1,7]	157	1054	948.2	7.2
‵				[4,1,1,8,1,1,8]	159	739	922.2	9.3
				[5,1,1,8,1,1,7]	157	1054	944.2	4.5
				[5,1,1,8,1,1,8]	159	739	928.7	6.5
				[6,1,1,8,1,1,7]	159	1054	949.4	8.1
				[7,1,1,8,1,1,7]	159	1054	952.4	7.1
				[8,1,1,8,1,1,7]	159	1054	955.3	7.5
				[7,1,1,6,1,3,7]	159	1159	922.3	12.7
				[8,1,1,6,1,3,7]	159	1159	927.4	6.2
	5	20,285	13	[2,1,1,8,1,1,7]	150	1053	963.0	5.4
				[2,1,1,8,1,1,8]	153	738	931.9	10.5
				[3,1,1,8,1,1,7]	153	1054	955.2	8.4
				[3,1,1,8,1,1,8]	155	738	931.1	6.5
				[4,1,1,8,1,1,7]	157	1054	942.9	9.6
				[4,1,1,8,1,1,8]	159	739	925.3	7.4
				[5,1,1,8,1,1,7]	157	1054	945.2	9.6
				[5,1,1,8,1,1,8]	159	739	925.0	10.6
				[6,1,1,8,1,1,7]	159	1054	948.9	4.1
				[7,1,1,8,1,1,7]	159	1054	956.0	5.8
				[8,1,1,8,1,1,7]	159	1054	951.2	7.9
				[7,1,1,6,1,3,7]	159	1159	922.2	8.9
				[8,1,1,6,1,3,7]	159	1159	921.7	17.6

Note:

^a ϕ and η denote threshold proportion in the core gene set and the candidate gene set.

^bSelection criteria: position j≤160, position l≤1200 and mean≥900. Definition of j and l is shown in footnote d and e below. The weight matrices with mean marked in bold.

^cWeight matrix is ordered by ω_association, ω_linkage, ω_{expression_human}, ω_{literature_human}, ω_kegg, ω_{expression_rat}, ω_{literature_animal}.

^dPosition j represents the position of the ϕ-th core gene locates in the η-th top ranked candidate genes.

^ePosition l represents the position of the last core gene locates in the ranked candidate genes.

^fMean: total number of random subsets having significant different p-value distribution from the top ranked candidate genes (Wilcoxon rank-sum test, p<0.05); sd: standard deviation.