Figure 1.

Structure of a hierarchical model. The replicates j of the generic subject i are normally distributed around a mean value μi with a within sample variance σ2, i.e. xij~N(μi, σ2). The mean values μi are normally distributed around a "population" mean value M with between sample variance τ2, i.e. μi~N(M, τ2).

Demichelis et al. BMC Bioinformatics 2006 7:514   doi:10.1186/1471-2105-7-514
Download authors' original image