CRII: III: Automated Discovery of Predictive Regulatory Models from Morphogenetic Experimental Data

University Of Maryland Baltimore County, Baltimore MD

Investigators

Abstract

Continuous advances in molecular biology are producing an extraordinary amount of detail about the necessary genes required for multicellular organisms to grow the correct shapes and forms during development and regeneration. However, reconstructing the genetic mechanisms explaining precisely the experimentally observed patterns and, more importantly, the morphological outcomes in terms of shape and forms, is an extraordinarily difficult task for human scientists due to the complex regulatory networks and dynamics characteristic of biological processes. This project will produce and use novel computational tools to aid in the discovery of dynamic predictive models explaining at once the exponentially growing experimental datasets of morphogenetic shapes, forms, and patterns. The interdisciplinary characteristic of this project will foster new links between computational and biological sciences, resulting in fundamental changes in the research methodology in these fields that will lead to an acceleration of novel applications in developmental and regenerative medicine. The computational tools developed in this work will also serve as educational advances for the teaching of dynamic regulation and systems theory in biology, computer science, engineering, and mathematics. Importantly, this project will involve graduate and undergraduate students from underrepresented groups in STEM fields. This work will be performed at UMBC, which has traditionally served large minority communities, and this project will advance the participation of these communities in the fast-paced field of computational biology. This research will develop the necessary set of computational algorithms and tools that will readily assist scientists in the discovery of the genetic mechanisms controlling the development of biological shapes, forms, and patterns. The technical methods to be employed will include: the use of a mathematical ontology for experimental perturbations, genetic expression patterns, and morphological outcomes; a computationally-efficient continuous modeling approach based on partial differential equations for both genetic regulation and biomechanical forces; and high-performance heuristic methods for the automation of the discovery of regulatory models. The project will be grounded on mathematical models of partial differential equations, which will leverage our understanding of biological mechanisms of morphogenesis using the theory of dynamical systems. Furthermore, these novel computational tools will allow us to streamline the building of predictive, dynamic models of morphogenesis, explain the ever-increasing large dataset of morphogenetic experiments being laboriously produced by the community, and advance many fields of biology and engineering.