The Pennsylvania Cancer Alliance Bioinformatics Consortium (PCABC) is a unique partnership comprised of the University of Pittsburgh Cancer Institute (UPCI), Fox Chase Cancer Center, Kimmel Cancer Center of Thomas Jefferson University, Penn State Cancer Institute, University of Pennsylvania Cancer Center and the Wistar Institute. The Consortium has the goal of using the state’s investment to cooperate closely to use a shared bioinformatics approach to identify biomarkers that have usefulness for assisting in the diagnosis of cancer and/or for predicting response to treatment of cancer patients or the clinical course of their disease.

The current project focuses on building a Data Warehouse that serves as a data access and mining platform for multiple data sets related to cancer research from various research centers. The data sets include Pennsylvania Cancer Registry, Clinical data, Tissue data, Biomarker, Geomarker and Gene Expression data. The goal is to integrate the distributed data sets mentioned above and offer a one-point access to all the participating research centers through a website.

top

Our research has lead us in exciting directions in which we have developed algorithms that help in identifying higher order domains and networks of labeled sites within the cell nucleus that are undergoing replication and transcription. We are extending this analysis to the more difficult problem of quantifying domain and network changes in living cells that undergo complex 3D motion with time. We have identified salient structural features that enable us to formulate this as a graph matching problem. Currently heuristic methods are under development that will improve the runtime and compute results with higher accuracy. In another research direction we have pursued, the chromatin has been characterized as a space filling fractal curve with a measurable fractal dimension. We have successfully applied this method to the classification of images of cell nuclei according to the stage of the cell cycle it was obtained in. We hope to extend this analysis to all stages of the cell cycle and also to distinguish between normal and cancerous cells.

top

DNA arrays measure expression levels of thousands of mRNAs in a single experiment. Data mining tools are required to be developed to mine this information rich data with the goal of identifying the patterns that characterize the underlying mechanisms of action. The kinetics of gene expression are commonly examined in many experimental designs to delineate the temporal sequence of transcriptional events that occur in response to a given stimulus. The identification of groups of genes with ‘similar’ temporal patterns of expression is usually a critical step in the analysis of kinetic data because it provides insights into the gene–gene interactions and thereby facilitates the testing and development of mechanistic models for the regulation of the underlying biological processes.

Our research work in this area has included the development of clustering algorithms specifically designed for temporal gene expression data. Information theoretic dissimilarity measures such as the Kullback-Leibler divergence for comparing the shapes of the two gene profiles have proven extremely useful.

top

Microarray experimental data are a valuable, but limited source for inferring gene regulation mechanisms on a genomic scale. Additional information such as promoter sequences of genes/ DNA binding motifs, gene ontologies, and location data, when combined with gene expression analysis can increase the statistical significance of the finding.

To this end, our current research work focuses on the development of a machine learning based algorithms for the combination of heterogeneous genomic data. Our current project involves the development of a general computational system that integrates diverse data from multiple genomic sensors such as gene expression data, upstream sequences of genes, genomic multi-species sequence alignment data, and putative regulatory regions, for example, in order to provide improved predictions of regulatory regions made genome-wide.

top

>>

top