I am interested in the problem of developmental regulation at the level of the binding of regulatory proteins to DNA binding sites. In particular, I am interested in computational methods for determining what subregions of DNA binding sites are essential for recognition and for the development of improved search algorithms for such binding sites. These methods include Shannon's local complexity measure, information content, modifications of the Berg-von Hippel statistical mechanical approach, 3-dimensional analysis of binding site topology, and back-propagation neural nets. Problems considered have included the CRP binding sites, promoter structure in Escherichia coli, and the recognition of tRNA's by synthetases microarray chip analysis. This work involves both the development of appropriate software and the application of that software in computer modelling studies.

Recently it has become possible to meld the statistical metrics with the neural network approach by using the modified Berg-von Hippel evaluation of a sequence rather than the sequences itself as input for the neural net. This produces a very effective general search engine.