TABLE 3.

Prediction with BLUP-filtered SNPs (in AUC).

Penalty*	Number of SNPs**	AA model AUC	EA model AUC	NHW model AUC
Ridge	0.1k	0.749 (0.0368)	0.786 (0.0244)	0.746 (0.0116)
	0.5k	0.743 (0.0391)	0.786 (0.0244)	0.743 (0.0155)
	1k	0.731 (0.0337)	0.786 (0.0244)	0.721 (0.0148)
	5k	0.739 (0.0363)	0.786 (0.0244)	0.751 (0.0113)
	10k	0.742 (0.0345)	0.783 (0.0266)	0.754 (0.0127)
	15k	0.741 (0.0359)	0.783 (0.0266)	0.753 (0.0105)
	20k	0.742 (0.0354)	0.783 (0.0265)	0.75 (0.0095)
Lasso	0.1k	0.712 (0.0342)	0.784 (0.0253)	0.735 (0.0132)
	0.5k	0.683 (0.0376)	0.784 (0.0253)	0.703 (0.0108)
	1k	0.681 (0.0403)	0.784 (0.0253)	0.679 (0.0128)
	5k	0.679 (0.03)	0.784 (0.0253)	0.671 (0.0061)
	10k	0.684 (0.0287)	0.784 (0.0255)	0.694 (0.011)
	15k	0.695 (0.0314)	0.784 (0.0255)	0.704 (0.0077)
	20k	0.7 (0.0323)	0.784 (0.0255)	0.699 (0.001)
Elastic net	0.1k	0.712 (0.0342)	0.784 (0.0253)	0.735 (0.0131)
	0.5k	0.683 (0.0376)	0.784 (0.0253)	0.703 (0.0108)
	1k	0.681 (0.0403)	0.784 (0.0253)	0.679 (0.0128)
	5k	0.679 (0.03)	0.784 (0.0253)	0.671 (0.0061)
	10k	0.684 (0.0287)	0.784 (0.0254)	0.694 (0.011)
	15k	0.695 (0.0314)	0.784 (0.0255)	0.704 (0.0077)
	20k	0.7 (0.0323)	0.784 (0.0255)	0.699 (0.001)

* All penalized models include Age, Sex, and Pack-years as covariates. ** SNPs with the largest absolute BLUP value were prioritized.