Regulatory Networks in Development: Understanding the role of cis-regulatory modules in Gene Regulatory Network evolution

Much of the diversity in the animal kingdom is achieved by alterations in the structure and function of Gene Regulatory Networks (GRNs). GRNs regulate the spatio-temporal gene expression patterns of developmental genes via cis-regulatory modules (CRMs) and thus form a critical component of GRNs. CRMs comprise the basic units of GRNs and serve as integrated hubs for transcription factors (TFs) to bind to and regulate the genes in the network. Annotation of CRM elements is thus critical to understanding how changes in CRMs or the TFs that bind to CRMs can drive the evolution of GRNs. However, identification of CRMs has been particularly challenging despite the availability of sequenced genomes for many species, especially nonmodel or emerging model organisms, owing to inherent difficulties in identifying them. We have extended previous computational methods for CRM detection to predict over 8000 CRMs in six evolutionary diverged species including Anopheles gambiae, Aedes aegypti, Tribolium castaneum, Apis mellifera, Nasonia vitripennis as well as in Drosophila melanogaster. Functional validation of over 50 of these predictions in transgenic Drosophila demonstrate the power of our prediction pipeline (called SCRMshaw ??Supervised CRM discovery) and its use in elucidating GRNs across divergent organisms at the level of CRMs and their interacting partners.;Using SCRMshaw, we have unraveled key differences in the regulation of several developmentally critical genes for central nervous system development in two highly diverged insect species Aedes aegypti and Drosophila melanogaster leading to neofunctionalization of an ancient conserved GRN in Ae. aegypti.;We have also continued our efforts to improve SCRMshaw with a new evaluation pipeline that now offers better measures to evaluate the sensitivity of SCRMshaw when employed for CRM prediction in species outside of D. melanogaster as well as enabling us to assign confidence scores to each predicted CRM in a given species.;Together, the work described here provides a general framework for the development, evaluation and use of computational CRM discovery methods towards understanding the evolution of complex traits as well as key developmental process in related and distant metazoan species.