Using a Bayesian latent variable approach to detect pleiotropy in the Genetic Analysis Workshop 18 data

BMC Proceedings

Table 3 Results of the Bayesian LV method applied to rs9816772, previously identified as associated with DBP

	Parameter	Estimate	logBF	95% HpdI
SBP	$λ_{1}$	13.15	255.3	(12.19, 14.11)
DBP	$λ_{2}$	7.60	139.6	(7.01, 8.14)
Sex for SBP	$β_{11}$	−0.66	−0.074	(−2.12, 0.81)
Sex for DBP	$β_{21}$	−1.79	2.017	(−2.92, -0.65)
rs9816772	$α_{1}$	−0.045	−0.653	(−0.208, 0.124)
Age	$α_{2}$	0.043	126.53	(0.036, 0.049)

$λ_{j}$ is the factor loading for the association between phenotype Y_j (j = 1, 2: SBP and DBP) and the conceptual latent variable U, $β_{j k}$ evaluates the association between phenotype Y_j and covariate W_k( k = 1: sex), $α_{k}$ captures the association between the latent variable U and covariate X_k (k = 1, 2: SNP and age).

ISSN: 1753-6561