Overview
With a PhD in theoretical physics and over twenty years experience in experimental biological research, it is not surprising that my current research is in computational biology, focused on the quantitative analysis of experimental data in biochemistry and cell biology. Similarly, I am particularly interested in the training of both our undergraduate and graduate biology students at the interface between biology, mathematics, physics and computation.
Research Focus
Our research focuses on the role of aberrant splicing of the Protein Tyrosine Phosphatase alpha (PTPalpha) proto-oncoprotein in cancer. Our previous experimental studies with a limited number of patients showed that PTPalpha is aberrantly spliced in large fractions of breast and colon cancers and that it can induce cancer by "dominant-positive" activity. We are now computationally investigating much larger numbers of patients and additional tumor types by analyzing high-throughput RNA sequence data from human tumors. In experimental studies, we are investigating the mechanism of alternative splicing and whether known cancer-associated mutations are acting via the dominant positive mechanism.
Additional projects in computational biology include
1) Mathematical modeling of the dynamic regulation of kinases and phosphatases during vaso-constriction and -dilation We have discovered a new mechanism of "regulation by unfair conservation" and are testing the hypothesis that it is employed in multiple physiological situations. (Collaboration with Michael Goldberg, Dept. of MBG.)
2) Optimization of SELEX with high-throughput sequencing. Systematic Evolution of Ligands by Exponential Enrichment is a powerful procedure for identifying, from starting pools containing astronomical numbers of molecular variants, small RNA molecules (aptamers) that bind to designated targets, often for use as probes or inhibitors. High-throughput sequencing and multiplexing techniques have recently been incorporated to improve the efficiency and performance of the method, but it can be improved further by carefully optimizing the experimental conditions that are used. We are using statistical and computational modeling of the procedure towards this end. Our previous studies have shown how the selection schedule and data clustering can be used to improve analysis and that non-specific cooperative binding of aptamers to immobilized targets can severely degrade performance and have shown how this effect can be minimized. We are currently studying how the selection schedule can be further optimized (Collaboration with John Lis, Dept. of MBG).
3) Cell replication and motility in the epidermis and hair follicle. These systems provide models of organ development that can be studied in detail using transgenic mice in which the replication and motility of individual cells can be followed using fluorescent tracers. We use mathematical modeling to infer the underlying biological processes involved. (Collaboration with Doina Tumbar, Dept. of MBG.)
4) Improved methods of pattern recognition in multidimensional large database analysis. Even the identification of patterns in a two-dimensional field (e.g., in vision) is computationally challenging, and the many-dimensional problem posed by large-scale bio-databases is even more difficult. We are developing a novel approach using advanced methods from stochastic statistical physics.
