Understanding the genetic mechanisms through which complex traits arise remains one of the most important and challenging problems in biology. There has been impressive progress in sequencing genomes, cataloguing genes and describing the biochemical functions of their products. However, the network of genetic and biochemical processes that form and change organisms, the "wiring diagram of life" is not well understood. There are two approaches to understanding biological networks. The first, or "bottom up" approach, starting at the molecular level, has to date had limited success. Geneticists, using a second or "top down" approach have had success in deducing the molecular mechanisms that underlie genetic traits. However, applying this top down method to complex traits is extremely hard because of the highly complicated systems that underlie the expression of such traits. This project will develop and test new mathematical and computational methods to infer, using top down approaches, the networks that produce complex traits. For the connections between traits and mechanisms a multidimensional mathematical function is substituted, allowing estimation of the probability that a proposed connection between a trait and a group of genes is correct. This mathematical function uses existing automated reasoning techniques to develop new applications to biology. For moving from the trait down to the network, models will be generated and evaluated using a set of rules that encodes the knowledge and experience of geneticists. Two model systems from corn (maize) will be used to test these methods. The first will utilize a small, very well understood set of biochemical pathways that determines the colors of corn kernels to test and validate the computational methods. The second is a system of approximately 200 genes that are involved in the plant's defenses to pathogens, including fungi that devastated the U. S. corn crop in the 1970s and that are a world-wide threat to corn crops. The goal of this second part is to characterize and accurately quantify the plant's responses to infection, using a set of mutant genes that mimic the plant's defensive response to pathogens, and to infer the underlying genetic network responsible for these responses.
Broader Impacts. Food, fuel, and climate security depend on improving crop yields while reducing the nutrients, agricultural chemicals, and water used by the farmer. This project will develop and test new tools for understanding how best to meet these challenges. The project brings together an interdisciplinary, multi-institutional team of researchers who will provide training at the undergraduate and postdoctoral levels in plant biology, computational biology, and biostatistics, a combination increasingly needed in both academic and commercial enterprises.