Estimating heritability of gene expression using parent-offspring regression with 2-channel microarrays

With the advent of microarrays, it is possible to look at the entire transcriptome of an organism as a suite of quantitative traits. An obvious question to now ask is: To what extent is gene expression heritable? In quantitative genetics, single parent-offspring regression is the most straightforward method in situations where the progenies are produced by cross-pollination to many male parents of unknown location. However, estimation of the heritability of gene expression with single parent-offspring regression has not yet been examined with 2-channel microarrays. Here we introduce 3 experimental designs: chain design, independent quartets design, and completely independent design. We then compare them with common reference design in respect to statistical power and bias of the estimates. In our simulations, we also incorporated a model of simple inheritance with one gene. The results of our simulations indicate the efficiency of the chain design over the alternative design considered.