Analysis of variances ANOVA

Applying ANOVA for microarray data analysis was first proposed by Kerr et al. (20). Assuming a dye-swap experiment setting, the underlying statistical model is: log(yijk) = | + At + Dj + Vk + Gg + (AG)g + (VG)kg + eijkg where | is the overall mean, At the overall array effect, Dj the overall dye effect, Vk the overall variety effect and Gg is the overall gene effect across the other factors. The (AG)ig term describes the potential effects which are specifically due to the variation in the amount of spotted cDNA on the array. (VG)kg describes signal intensities explained by the considered variety (probe vs control), which is the main factor of interest in detecting biological differences. The error term eijkg is assumed to be normally distributed around zero. Parameters are calculated by maximum likelihood estimation. The authors provide codes for Matlab and R on their homepage:

0 0

Post a comment