The big picture
A recent analysis of Medline and Crisp showed that the top 2000 human and mouse genes are disproportionately represented in the literature and funded grants (Su et al., Genome Biology, 2007). While this trend has resulted in broad and deep knowledge for these key regulators and their pathways (e.g. P53, Vegf, NFkB), the social tendency of researchers to study previously studied and funded genes has had a negative impact on whole genome annotation. The past decade has seen several technologies (proteomics, RNA expression dynamics, cell based screening) and paradigms (functional genomics, systems biology) begin to shift the pendulum from one gene at a time towards whole genome research. Our laboratory is developing and utilizing these approaches to study mammalian pathways such as the circadian clock.
Systems biology: a path to the understanding of clock properties
For the last several years, our laboratory has begun to apply systems biology approaches to better understand complex mammalian behavior and physiology. Systems biology has been described as the antithesis of reductionism, and can be thought of as a simple process: identifying the components of a system, determining how those components fit together, and developing models that describe the emergent properties of the system (e.g. behavior) (Sauer et al., Science, 2007). While the reductionist approach has identified many components of biological pathways (and determined many key interactions), it has been less successful in describing how these interactions culminate in the emergent properties of systems. Practically, the process of systems biology can be accomplished by perturbing a system, determining the effects of these perturbations in a rigorous and broad manner, and incorporating this information in robust computational models that both describe the observed results and lead to testable predictions on the behavior of the system. Although important inroads have been made in realizing these goals in prokaryotes and other model systems, a systems level model of a mammalian behavior has remained elusive.
Why study the clock with systems biology approaches? First, the sleep wake cycle and dependent physiology and behavior in mammals emerge from interactions of (more) simple cell autonomous oscillators, or clocks that run in single cells (Ko et al., Hum Mol Genet., 2006). These rhythms are highly robust and can be studied after perturbation in a multi-parametric fashion. For example, rhythms in locomotor activity and temperature can be measured in the mouse for months at a time resulting in reliable detection of a ten minute difference in period length. These rhythms can also be observed in peripheral organs and even in immortalized fibroblasts and tumor cell lines (Balsalobre et al., Cell, 1998), where they can be easily perturbed with small molecules or by gene dosage experiments using over expression or RNA interference. In addition, these phenotypes can be measured at the level of the single cellâ€”indeed single cell measurements are informing our understanding of how the circadian network oscillator regulates behavior and physiology (Sato et al., Nature Genetics, 2006; Liu et al., Cell, 2007). But maybe the most important reason we are moving forward with the systems biology approach is that reductionist biology fails to explain basic properties of the clock such as robustness, temperature compensation, or periodicity, which are emergent properties of networks, and not the products of single genes.
Understanding the function of noncoding RNA
One of the most exciting findings of the past decade has been the discovery of novel roles for noncoding RNA. These RNAs have been shown to play roles in many if not most cellular processes such as transcription, splicing, translation, regulation of protein function, And as regulatory sensors of small molecules (Amaral et al., Science, 2008). Large-scale cDNA sequencing experiments like those at Riken and tiling array experiments on the Affymetrix platform have shown that the majority of the genome is transcribed (Cawley et al., Cell, 2004; Carninci et al., Science, 2005). Understanding the function of these noncoding RNAs has proven to be a challenge. Pioneering work has uncovered the roles of several classes, for example, micro-RNAs that regulate both translation and transcription (Lee et al., Cell, 1993; Ambros, Cell, 2004). However, determining the cellular roles for these factors at scale has proven more difficult. Computational tools have been developed to allow one to predict targets of noncoding RNAs such as micro-RNAs. These predictions, though, indicate potential regulation, which may occur in a tissue specific fashion and may or may not be relevant to cellular function. Our laboratory is using cell based screening to pan libraries of short hairpin RNAs and siRNAs designed against noncoding RNAs, as well as synthetic micro-RNAs and antagomirs to identify functional noncoding RNAs in signal transduction pathways (e.g. Willingham et al., Science, 2005). These screens are being integrated with other genomic screens as well is biochemical and computational approaches to identify those noncoding RNAs that play important roles in regulating mammalian biology.