Systematic determination of genetic network architecture

Technologies to measure whole-genome mRNA abundances(1-3) and methods to or ganize and display such data(4-10) are emerging as valuable tools for syste ms-level exploration of transcriptional regulatory networks. For instance, it has been shown that mRNA data from 118 genes, measured at several time p oints in the developing hindbrain of mice, can be hierarchically clustered into various patterns (or 'waves') whose members tend to participate in com mon processes(5). We have previously shown that hierarchical clustering can group together genes whose cis-regulatory elements are bound by the same p roteins in vivo(6). Hierarchical clustering has also been used to organize genes into hierarchical dendograms on the basis of their expression across multiple growth conditions(7). The application of Fourier analysis to synch ronized yeast mRNA expression data has identified cell-cycle periodic genes , many of which have expected cis-regulatory elements(8). Here we apply a s ystematic set of statistical algorithms, based on whole-genome mRNA data, p artitional clustering and motif discovery, to identify transcriptional regu latory sub-networks in yeast-without any a priori knowledge of their struct ure or any assumptions about their dynamics. This approach uncovered new re gulons (sets of co-regulated genes) and their putative cis-regulatory eleme nts. We used statistical characterization of known regulons and motifs to d erive criteria by which we infer the biological significance of newly disco vered regulons and motifs. Our approach holds promise for the rapid elucida tion of genetic network architecture in sequenced organisms in which little biology is known.