Unlike most bioinformatics data, such as DNA sequences and protein structures, pathways show the interactions of different bio-chemical entities. Thus, each entity or subpathway can have a function in its pathway that the pathway can not realize without it. Formulating these functions presents exciting computational challenges.
This proposal develops algorithms for efficient and accurate computational analysis of pathways. The first step in achieving this goal is to build a mathematical model for functions of pathways and subpathways. This proposal defines the function of a subpathway as its contribution to the steady state of the entire pathway. Computing this contribution is a difficult problem especially for gene regulatory pathways. Existing methods can compute this for metabolic pathways, but they do not scale to even medium sized gene regulatory pathways. This proposal develops efficient methods for computing function.
Finding the subpathway that has a desired function is a challenging problem. There is no clear way of searching for subpathways with desired functions using existing tools. This proposal develops efficient algorithms for searching subpathways with desired function. This proposal takes this one step further, and develops mathematically sound algorithms for comparing two pathways so that the entities that map are functionally, homologically and topologically similar. This proposal also explores feature and reference-based index structures for biological pathway databases.
Further information on the project can be found at the project web page: http://bioinformatics.cise.ufl.edu/
