Mixture SNPs effect on phenotype in genome-wide association studies

BMC Genomics

Table 4 Real data estimation results using HBM and MLM

	The Framingham heart study^a		The health and retirement study^b
	Estimates from	Estimates from	Estimates from	Estimates from
Parameter	HBM (s.e)^b	MLM (s.e)^b	HBM (s.e)^b	MLM (s.e)^b
(Intercept)	26.42 (0.29)	26.46 (0.28)	27.47 (0.07)	27.47 (0.05)
	0.20 (0.09)	0.014 (0.005)	0.014 (0.009)	0.0001 (0.00002)
	24.68 (0.38)	22.64 (0.41)	22.32 (0.23)	27.29 (0.41)
	1.49 (0.40)	3.49 (0.35)	0.6678 (0.4)	0.95 (0.38)
	0.0034 (0.0005) ^c	-	0.0042 (0.0004) ^c	-
PVE	0.06 (0.01) ^d	0.13 (0.01) ^d	0.026 (0.01)	0.04 (0.01)
PC1	-19.01 (44.72)	-10.90 (44.15)	82.75 (33.56)	91.66 (20.10)
PC2	-18.00 (24.41)	-18.27 (24.90)	-15.54 (8.52)	-18.41 (10.49)
PC3	2.63 (8.91)	5.65 (10.25)	26.01 (9.32)	23.01 (6.91)
PC4	-10.13 (9.11)	-10.05 (9.45)	3.52 (3.24)	2.24 (6.83)
PC5	18.55 (9.44)	19.80 (9.89)	14.43 (11.14)	8.49 (6.79)
PC6	-4.75 (12.17)	-4.49 (12.02)	-9.77 (14.34)	-19.81 (6.77)
PC7	16.06 (12.46)	13.09 (12.48)	-1.21 (11.95)	-2.39 (6.74)

^aThe analysis is based on 1,915 unrelated persons in the Framingham Heart data set using 13,764 SNPs on Chromosome 16 to predict BMI. ^bThe analysis is based on 12,237 unrelated persons in the Health and Retirement Study using 11,925 SNPs on Chromosome 16 to predict BMI. ^cValues in the parenthesis are standard errors. ^dPVE estimated by MLM is higher than that estimated by HBM.

ISSN: 1471-2164