"Our current research combines three decades of expertise in biotechnology and artificial intelligence. This research has resulted in nearly a hundred publications that have been cited over a thousand times. The research has also resulted in the training of nearly 40 graduate students and post-doctoral research fellows (who are currently well placed in industry and academia) and about 50 undergraduate research students. The key research accomplishments are in the fields of Biochemical engineering and "On-line fault diagnosis". Current research is in the area of Mathematical/Computational/Systems Biology. In biotechnology, we are currently applying mathematical models and knowledge-based approaches to convert data to useful knowledge. In chemical process systems, we are analyzing on-line data using hybrid mathematical-heuristic approaches to diagnose faults. Common to both these application areas are the use of dynamic models, qualitative domain knowledge and artificial intelligence approaches for data interpretation and knowledge integration.
Our past research in biotechnology has resulted in a better understanding of microscopic and macroscopic variables that influence the kinetics of genetically engineered microorganisms. The work from 1977 to 2000 in biochemical engineering was largely experimental. In joint research with the Pasteur Institute in Paris, Kostas Tokatlidis examined the effect of gene sequence and protein structure on the mechanism of inclusion body (protein aggregate) formation during high-level expression of cellulolytic proteins in recombinant cells. As part of collaborative research with DuPont in environmental biotechnology, Konstantin Konstantinov (DuPont/Delaware) investigated the use of bioluminescent recombinant microorganisms as sensors for pollutants. Our paper on bioluminescence has been cited over 150 times and the paper on cellulase-based inclusion bodies has been cited over 100 times.
A second major area of our research is Knowledge-Based Process Monitoring and Fault Diagnosis. In a joint project with Foxboro and DuPont (1982-86), we were involved in a pioneering effort for the first industrial application of an expert system, FALCON (Fault Analysis Consultant), for on-line fault diagnosis in a dynamic chemical process (DuPont adipic acid plant in Victoria, Texas). A key aspect of our work with intelligent monitoring and diagnostic systems is the exploitation of qualitative domain knowledge from heuristics as well as quantitative knowledge from mathematical models.
We are currently combining our past expertise in biotechnology and knowledge-based systems to examine novel applications in Bioinformatics and Systems Biology. The introduction of a gene into an organism perturbs it at many levels of cellular hierarchy. Genomic, Transcriptomic, Proteomic and Metabolomic data provide insight into the effects of such manipulations. We are interested in modeling these effects and interpreting the large amounts of data using mathematical models and qualitative knowledge. We have developed a simulator, eXPatGen, www.che.udel.edu/eXPatGen capable of simulating dynamic gene expression profiles resulting from complex regulatory interactions. This simulator can be used to evaluate multiple ways of analyzing the expression profiles and determining the best methods of analysis. In a joint project with Allan Shapiro in Plant Science, we have developed a mathematical model of signaling pathways in Arabidopsis disease resistance. The model has been used to provide hypothesis and guidance for new experiments. The goal is to develop system level language and tools for analysis and integration of information at various levels of hierarchy in biological systems. Our recent review article on Systems Biology (April 2008) points to the future direction of our research."
Selected PublicationsSelected Awards
