Seralogix, inc. is proposing to develop and validate new computational tools and methods in conjunction with an intelligent framework for the identification, analysis, and modeling of the mechanisms and pathways associated with the host-pathogen innate immune and inflammatory responses to biowarfare infectious agents. Our core computational tool is based on the statistical power of dynamic Bayesian networks (DBNs), which is utilized to model the complex dynamic pattern-of-change of DNA, mRNA, proteins and metabolites, which we refer to as the """"""""temporal biosignature"""""""" of the host-pathogen response. DBNs are based on sound statistical methods that allow us to combine prior knowledge with time-course empirical data for deciphering host-pathogen biosignatures with a biological system perspective. Because of the complex multi-dimensional data resulting from genomic and proteomic investigations, new computational tools, with built in intelligence, are required to serve the investigative needs of the 21st century. Our approach utilizes the power of """"""""intelligent software agents"""""""" to create an integrated framework that automates: 1) steps in our DBN methodology, 2) importing and managing complex genomic and proteomic data, 3) the continuous collection and updating of prior knowledge from other sources, 4) the intelligent and error correcting construction and modification of DBN models, 5) the guided selection and application of appropriate computational techniques and analytical algorithms, 6) the visualization of results, and 7) the implementation of a self-learning process. A unique feature of our approach is the added dimensionality of """"""""time"""""""" combined with prior knowledge that we believe enables new host-pathogen time-course investigations that hold the promise of deciphering the causal relationships across the intracellular and intercellular signaling domains. We believe that our methods and tools will aid in the identification of new targets/pathways of intervention and statistically confirm biological impact of intervening drugs and therapeutic treatments. We hypothesize that our DBN methodology should substantially improve the statistical significance for inferring innate and inflammatory pathways from less experimental data while also confronting noisy, hidden, and/or missing data points. The Phase I goal is to develop a """"""""beta"""""""" version of our computational tools and framework to demonstrate feasibility and merit. The Team includes Seralogix Jnc, the University of Texas Medical Branch Department of Microbiology and Immunology and the Center for Biodefense, Galveston; the University of Texas Center for Biomedical inventions (UTSWCBI), Dallas; and the Walter Reed Army institute of Research.
