BMOL Seminar Series: Dr. Tom Glass

Seminar Details

Speaker: Dr. Tom Glass, Senior Scientist at Sapidyne Instruments Inc.

Host: Maddie Grier, BMOL Ph.D. Student

Title: Exploring KinExA® Technology and Critical Analysis of 95% Confidence Intervals for Fitted Parameters

Abstract: This presentation will explore the fundamental principles of the KinExA method for determining binding constants which consistently provides 95% confidence intervals for fitted parameters. We will discuss the methodology used for estimating these confidence intervals and highlight its advantages over competing approaches. Through critical analysis, we will examine how confidence intervals enhance data interpretation and decision-making in binding studies. Audience participation and questions will be encouraged throughout the session, fostering an engaging discussion on the nuances of KinExA technology and statistical analysis.

Publications Related to Talk:

This book provides a concise, easily readable introduction to topics related to regression analysis and in Chapter 19 describes use of the F distribution for estimation of parameter confidence regions. This technique is closely related to the one used in KinExA.

• Fitting Models to Biological Data using Linear and Nonlinear Regression

This paper describes the importance of asymmetric confidence intervals for binding constants from the perspective of calorimetry measurements. Confidence intervals for binding constants are inherently asymmetric and reporting symmetric confidence intervals is never appropriate.

• Uncertainty in protein–ligand binding constants asymmetric confidence intervals versus standard errors

From the perspective of enzyme kinetics Johnson et al. discuss the exploration of the error surface as a means of estimating parameter confidence intervals including the possibility and importance of unbounded intervals in cases of unconstrained parameters. In my talk, I’ll also discuss the critical importance of parameter constraint and how parameter constraint leads naturally to asymmetry of the corresponding confidence intervals.

• FitSpace Explorer: An algorithm to evaluate multidimensional parameter space in fitting kinetic data