6 . 047 / 6 . 878 Fall 2008 Lecture 24 : Module Networks

Biological systems are generally incredibly complex, consisting of a huge number of interacting parts. Consider, for example, the task of understanding the structure of the genetic regulatory system. The dependence among expression levels of all the genes in a cell is so complicated that it makes any sort of detailed understanding almost impossible. The diagram for the functional network of yeast genes looks more like a hairball than anything else! Therefore, for a high-level, conceptual understanding of how transcription is controlled, it is useful to talk about higher-level objects rather than individual gene expressions. Often such conceptually simpler networks are drawn by hand by biologists. Is there a systematic way to accomplish this? To this end, we de ne the notion of a module, which is a set of biological entities that act collectively to perform an identi able and distinct function. A module should act as an entity with only weak linkage to the rest of the system. Some examples of modules are metabolic pathways, molecular machines and complexes for the function of signal transduction. Our focus here is on regulatory modules, modules in the genome that perform a distinct function. Regulatory modules are co-expressed, co-evolved and regulated by the same set of transcription factors. For example, an operon, which consists of genes coded next to each other and constitutes a functional unit of transcription, is a regulatory module. Another paradigmatic example is the ribosomal module. Using modules as the basic units, one could now construct a regulatory module network where each node is a regulatory module and where interdependencies between modules are well-de ned since all the genes in a module are co-expressed and co-regulated. The module network captures redundancies and structure not easily represented in a complete regulatory network. Now, module networks can be described in several ways. They can de described in terms of their functional behavior as a whole; this is the simplest approach but sometimes su ces. Or it can be described as a logical program. This is the approach we will take in the following and describe in detail. Or it can described in terms of the low-level chemistry that determines the dynamics of the network. This is hardest to accomplish but sometimes necessary.