Check this box to perform
principal component
regression where
principal components
(PCs) are calculated from all
SNPs
(or collapsed rare variant loci) with the same value of the
Annotation Analysis Group Variable
, and the PCs with the largest
eigenvalues
are included in a
regression
model similar to the multiple-
locus
regression model (with SNP
genotype
replaced by the PCs) (Gauderman et al. 2007, Wang and Abbot 2008). The number of PCs to include in the
model
is determined by the
Maximum Number of Principal Components
and
Cumulative Proportion of Variation to Explain with Principal Components
parameters below. You can adjust for
covariates
or class effects by specifying them on the
Model Variables
tab. Statistics for the joint test of all PCs included in the model are reported.
