• July 19, 2007 - Max-Planck Institute, Heidelberg, Germany (Host: Dr. Jochen Reinstein, Molecular Chaperones Group)
• July 17, 2007 - University of Regensburg, Germany (Host: Dr. Sandra Schlee, Department of Biochemistry)
• July 5, 2007 - University of Limerick, Ireland (Host: Dr. Jakki Cooney, Department of Chemical & Environmental Sciences)
• Download PowerPoint Slide Show (11.3 MB)
• Download PDF Handout (1.40 MB)

ABSTRACT

Biochemical kinetics is useful in the study of molecular mechanisms, but two serious challenges remain. First, how do we know that the fit of any given theoretical model (reaction mechanism) to our experimental data truly is the best available? Could another combination of kinetic constants provide a better fit? Second, how do we know that any given mechanism truly is the best available model, given that many possible mechanisms exist even for relatively simple biochemical reactions? Answers to these questions are offered by a computational technique called Differential Evolution (DE), which mimics the Darwinian process of random mutation and natural selection. We start with a "supermodel" -- the largest reasonable mechanism, possibly containing redundant steps. In the process of evolutionary optimization ("supermodel evolution"), any unnecessary branches in the reaction mechanism are then whittled away.

• March 8, 2007 - Washington State University (Host: Dr. Ron Brossemer, School of Molecular Biosciences)
• November 17, 2006 - Oregon Health Sciences University / OGI (Host: Dr. Jim Wittaker, Department of Environmental and Biomolecular Systems, School of Science & Engineering),
• Download PowerPoint Slide Show (9,204 kB)
• Download PDF Handout (815 kB)

ABSTRACT

Traditional enzyme kinetics relies on tediously deriving complicated mathematical equations. A relatively new approach lets the computer do all the algebra, so we can focus on the underlying biochemical mechanisms. For example, instead of typing v = Vmax [S] / ([S] + Km) into a data fitting program (the Michaelis-Menten equation), we can just type E + S <==> ES --> E + P (the Michaelis-Menten mechanism). I will describe an application of this "math-free" approach to enzyme kinetics, to determine the molecular mechanism by which the Lethal Factor protease from Bacillus anthracis (anthrax) is inhibited by aminoglycoside antibiotics.

• September 15, 2004 - Society for Biomolecular Screening, Orlando, Florida (Hosts: Dr. Michael Snowden and Dr. Ji-Hu Zhang)
• Download PowerPoint Slide Show (282 KB)
• Download PDF Handout (259 KB)

Talk given at the 10th Annual Society for Biomolecular Screening conference in Orlando, Florida, September 11-15, 2004. This was the last lecture in Session 5C, "Assessing and Maintaining Quality in HTS Data" (Michael Snowden & Ji-Hu Zhang, session co-chairs).

ABSTRACT

IC₅₀ is the most widely used measure of potency in secondary dose-response screening of enzyme inhibitors. However, IC₅₀ is not sufficiently informative if an inhibitor happens to be 'tight binding'. In this case, a more appropriate measure of potency is the apparent inhibition constant. We have developed new theoretical methods for fully automatic determination of apparent inhibition constants from dose response data [P. Kuzmic et al. (2000) Anal. Biochem. 281, 62-67] and recently described new theoretical approaches to data-quality management in secondary screening projects. For example, automatic outlier rejection can be achieved by a modification of Huber's Minimax method [P. Kuzmic et al. (2004) Meth. Enzymol. 383, 366-381]. This presentation will focus on issues that arise in successfully implementing the underlying kinetic theory and advanced statistical methods (robust regression, asymmetrical confidence interval estimation) in data-analytical practice. The results will be illustrated on a typical secondary screening project involving 10,000+ inhibitors.