Wendy Warr interviews Anton "Tony" Hopfinger

Professor Hopfinger gained his PhD at Case Western Reserve University in 1969.

He is currently Distinguished Research Professor of Pharmacy at the University of New Mexico and Professor Emeritus of Medicinal Chemistry in the College of Pharmacy at the University of Illinois at Chicago.

His current research interests are 4D-QSAR analysis to develop structure-activity hypotheses for designing new drugs; prediction of ligand-receptor binding thermodynamics; membrane-interaction QSAR analysis for ADMET property prediction; and characterization of molecular similarity and virtual screening. He is an Associate Editor of the ACS Journal of Chemical information and Modeling.

Download PDF Version

Dr. Warr: What attracted you to QSAR research in the first place?
Prof. Hopfinger: I didn’t start out in QSAR research, but rather in developing methods of doing conformational analyses of molecules, including synthetic macromolecules. As my interest grew in medicinal chemistry and pharmacology, the identification of the "active" conformation of a drug from a conformational analysis became of high interest to me. However, I soon came to appreciate what medicinal chemists and pharmacologists already knew. The active conformation is not sufficient to understand and model most structure-activity relationships. Hence, my interests then turned to seeking how I could represent conformational information within the Hansch QSAR paradigm. And that was the start of my interest in QSAR, as well as the start of one of the initial efforts in developing the 3D-QSAR field.

W: I think that I probably first heard your name when MDL was marketing your Chemlab software back in the 1980s. Things have moved on since then, haven’t they?
H: Chemlab is probably one of the first two molecular modeling packages to be commercialized in the USA. Interestingly, Chemlab had its origins in providing a way to capture and reuse software developed and/or used in my lab by my graduate students and postdocs. In a sense, I didn’t want to "reinvent the wheel" in terms of software each time I got a new student. Today, the choice of software products for use in solving chemical problems is mind boggling in both number and diverse functionality. Unfortunately, and this is solely my subjective view, software products have become so numerous, and with such a major focus on ease of use, that many users are not really aware of what they are using as long it runs for their study. That is, there seems to be less of a concern knowing the hard science behind a calculation now than in the past.

W: You are famous for 4D-QSAR but 5D and even 6D QSAR are now reported. These terms are confusing to cheminformaticians used to talking of 2D and 3D structures. Explain to me in words of one syllable what nD QSAR is.
H: I don’t know about "famous" but my group and I have developed a paradigm we call 4D-QSAR that somewhat surprisingly to us can be used in an identical and comparable manner on informatics-based, ligand-based and structure-based QSAR problems. The choice in name "4D-QSAR" may not, in retrospect, have been a good one. The fourth "dimension" in our paradigm is sampling and includes the sampling of conformation, alignment, pharmacophore sites and entropy. The composite information coming from each of these sampled property sets is embedded in the resulting QSAR model. Most of the other nD-QSAR methods consider each of these properties, which we sample as individual dimensions in their QSAR studies. Hence, by way of examples, one would get a 5D-QSAR method if conformational sampling is included, and a 6D- QSAR approach if both conformational and alignment samplings are considered.