One of the main challenges in the discovery of intracellular biomarkers and identification of therapeutic targets is the lack of a mechanistic understanding of the complex underlying pathways. The tremendous increase in both the quantity and diversity of cellular data represents a significant challenge to researchers seeking to construct biologically relevant interaction maps, and objectively extract specific actionable information. Machine learning based clustering algorithms serve as a preliminary statistical data analysis metric, but they fail to capture the data in the proper biological context. While chemical kinetics based models have proved to be effective in elucidating the pathway mechanisms, accurate estimates for the model parameters are severely lacking and are often impossible to obtain owing to the inherent difficulties involved in making dynamic measurements of specific intracellular phenomena. Additionally, methods for rational prioritization and selection of critical intracellular interactions (in the absence of kinetic information) are sorely lacking. Therefore, there is a clear need for innovative software tools that enable quantitative analysis of available microarray data in a biological pathway context, ultimately leading to the objective identification of critical biological interactions, providing a direction for more focused future efforts. We propose to address this challenge by developing an automated software platform that utilizes microarray data to select and merge relevant canonical biological pathway models thereby placing significantly expressed genes in their biological context. The analysis software will utilize a microarray expression-weighted metric to objectively rank the most critical interactions within the network model using a novel chemical kinetics-free Boolean dynamics algorithm. In the Phase I effort, we will develop a software tool composed of an R library that enables the automated generation of a pathway model from a given microarray dataset. Additionally, a methodology, and associated R library will be developed to objectively rank critical interactions in the pathway model, using a microarray data expression-weighted metric. Demonstration and validation of proposed algorithm will be carried out using a well characterized lipopolysaccharide (LPS) stimulated RAW 264.7 macrophage system. In Phase II, we will extend the scope of the algorithmic framework to include proteomic and metabolomic weighting in the objective ranking of critical interactions, and add workflow improvements through the addition of a graphical user interface (GUI). Experimental verification and validation of critical interactions identified in Phase I will be carried out using gene-silencing techniques. We also intend to establish collaborative partnerships with commercial entities. The proposing team has extensive experience in the areas of systems biology and bioinformatics (CFDRC) and microarray data analysis (Shawn Levy, University of Vanderbilt). CFDRC has a strong track record in the commercialization of software and hardware.
Recently, there has been a tremendous increase in both the amount and diversity of cellular data available to researchers, representing a clear need for the development of advanced computational analysis software to enable the discovery of biomarkers of disease states, and identification of new therapeutic targets. However, currently available analysis tools do not consider the data in a proper biological context. This research proposes to develop an automated software platform that utilizes available data to develop and analyze mathematical models of complex processes in an automated fashion, resulting in the identification of critical intracellular processes. ? ? ? ?
