Table 4.

Results for both datasets using GRLR

Dataset	Chr	Selected Subset of SNPs	Test Statistic	Empirical P-value
heroin addiction	1	{rs1408830, rs965972}	1.26 × 10-07	0.027
AMD HK	10	{rs2736911, rs10490924, rs763720}	1.32 × 10-11	0.000

Analysis results for two published datasets using our GRLR method. Genomic search regions are constructed based on the SNPs whose single-locus p-values are no larger than a fixed threshold 0.05 for the heroin addiction data, and 0.01 for the AMD Hong Kong data; the truncation threshold within each search region is set to 0.05; the maximal length on each side of the central SNP is set to 10; the tuning parameter λ in GRLR model is set to 1.0. Empirical p-values are derived from 1000 permutations.