Current and Past Projects in the MEL group

POSITIONS AVAILABLE: postdoctoral, graduate, undergraduate

CANCER CHEMOTHERAPY IMPROVEMENT: Several of our projects are aimed at improving cancer chemotherapy drugs. We use the above in-house software plus commercial software to discover quantitative structure-activity relationships (QSAR) and and then aim at creating focused combinatorial libraries to explore the region of chemical space around the best lead compounds.

INVENTON: The original project began was started by Wipke, Pitman, and Koehler in the early nineties. Just about every graduate student since INVENTON's inception have contributed to the project. The main emphasis of INVENTON is de novo design of small molecules to meet design goals specified by chemists in the "problem description". Our goal was to bring concepts from mechanical engineering to chemistry. INVENTON can design inhibitors for proteins, or mimics of lead compounds, or just interesting molecules such as a self-assembling nanotube.

QSDOCK: This project began approximately six years ago by Wipke and Goldman (see publications). QSDOCK is designed to be a fast docking program with a different premise than most other widely available docking programs. QSDOCK is based on shape complementarity between the ligand and the active site, not chemical. QSDOCK uses "quadratic shape descriptors" to describe both the shape of the small molecule and the shape of the active-site. Docking is performed by aligning similar shape patches on the small molecule with similar shape patches in the active-site.

HIGH-THROUGHPUT SCREENING: This project is in collaboration with the Affymax corporation and is based on the Affymax developed three-dimensional fingerprint. Affymax has applied this fingerprint in QSAR studies of small molecules. In conjuction with the production of small molecule 3D fingerprints, our contribution is to produce 3D fingerprints of the active-sites of selected proteins. In order to produce active-site fingerprints, sitepoints are placed in the active-site to mimic the binding of virtual inhibitors. The protein 3D fingerprints are developed upon these sitepoints. These 3D active-site fingerprints, in turn, can be used to find potential small molecule inhibitors by scoring of the protein fingerprint with small molecule. We have also utilized 2D fingerprints.

ASITE: A program which will automatically find and perceive active-sites within proteins.The program is loosely based loosely on the cellular logic method of Delaney in the early nineties. We have made a signficant algorithm improvement to the cellular logic operations as well as addressed some of the key shortcomings of Delaney's original method. We have applied ASITE to our in-house database currently containing 1,200 protein crystal structures.

TUTORG: This project is aimed at developing an intelligent tutor system for teaching organic chemistry. In contrast to other tutoring programs in chemistry, this one actually understands chemistry! Brilliant undergraduates have helped in developing this program.