Table 2. Mean computation time (in hours) of various methods for risk prediction in WTCCC.

	BayesR		BSLMM	LMM	GPRS
Trait	All SNPs	500 SNPs
T1D	17.76 (0.19)	5.94 (1.19)	8.59 (0.63)	0.24 (0.06)	256.5 (10.07)
BD	18.40 (0.17)	3.33 (0.65)	10.27 (3.66)	0.25 (0.02)	202.9 (36.52)

Values are means with standard deviations in parentheses. Computations were performed using a computing cluster with Intel E5-2670 processors with 2.66 GHz CPUs. Shown is the time taken to analyze 10 replicates of the 80/20 split data for BD and T1D. BayesR (All SNPs) samples each SNP effect in each MCMC iteration. BayesR (500 SNPs) uses a modified sampling scheme.