June 9, 2009 - Dr. Dongxiao Zhu was awarded an NIH/NLM Exploratory/Developmental Research Grant (R21) for developing a novel informatics paradigm for inferring Gene Regulatory Networks (GRNs) in human disease. The $409,531 grant is two-year in duration and funds two PhD students and one postdoctoral researcher working in the frontier of bioinformatics and computational biology. Computational modeling is one promising approach to reconstruct GRNs from the increasingly available gene expression data and prior knowledge. However, it has been an ill-posed problem because the network topology is difficult, if not impossible, to directly observe in a single experiment, and high experimental costs prohibit collecting sufficient samples in a single study. As a result, current research attempts to construct a phenomenological graph to emulate some key features of the true GRNs, often relying upon overly stringent, sometimes unrealistic computational and/or statistical assumptions. Approaches that exploit biological prior knowledge are typically coupled with relaxed mathematical and/or computational assumptions. However, the trade-off is that they are often sensitive to errors in prior knowledge. We lack realistic models for high throughput data analysis that balance the weights of prior knowledge and experimental (data) evidence to predict GRNs in human disease. Due to these challenges, the development of predictive models lags far behind the pace of data accumulation.

This proposed research will implement an innovative two-stage predictive modeling approach to reconstruct GRNs that will overcome these challenges. It aims to decipher the key gene regulatory mechanisms for human disease development and progression. The implementation of this project is expected to significantly impact the research field of gene network modeling and will lay a solid foundation for further theoretical and algorithmic developments in this field. The PI will closely work with the two co-investigators at Tulane Medical Center and McGill University in executing this project. For more information about Dr. Zhu’s research, please visit: http://cs.uno.edu/~dzhu/