PROJECT SUMMARY (See instructions): In this proposal, we develop new interpretable statistical methodology to advance medical genetic, which seeks to understand the genetic basis of complex diseases and to identify disease-causing gene mutations. Since genes typically act through modules and pathways in modulating cellular functions, a standard approach in medical genetics is to follow candidate gene identification from patient data with an analysis of the network formed by interactions of these candidate genes. The question of how to best utililize gene-gene interaction (GGI) network information to uncover disease genes constitutes an urgent bottleneck in medical genetics research, and our goal is to provide a statistical approach to this problem. Our proposed research is to model gene-gene interactions as a Markovian network, which takes into account the latent growth process of a network. Markovian models are especially suitable for GGI networks because new gene interactions arise through evolution over time. We propose four distinct but highly inter-related projects: (1) we propose a general Markovian network model for GGI networks that encompass various existing models as special cases; (2) we propose a flexible method based on the repro samples framework to construct confidence sets for the central root nodes of a Markovian network, which can be used to identify disease genes on GGI networks with rigorous frequentist significance guarantees, a task that was not possible or easily achieved in the past; (3) We propose an innovative development to incorporate community/module structure into Markovian networks and devise novel methods to estimate the functional modules and hidden gene nodes of a given GGI network; and (4) we give concrete plans to validate our methods and any empirical findings we obtain using criteria common in bioinformatics and laboratory experiments. Our research agenda has the potential to transform both medical genetics and foundational statistical research. The Pl team comprises statisticians and geneticist with expertise in network analysis, statistical inference, medical genetics, machine learning and it is ideally suitted to conduct the proposed research. Successful completion of this research is expected to introduce and validate a new and powerful strategy for understanding the genetic etiology of complex diseases. The proposed work will also greatly expand the reach of statistical inference and uncertainty quantification, and fundamentally impact our approach of making inference for many problems in data science.