Wendy Warr interviews Anton "Tony" Hopfinger

W: I note that you review for a huge number of journals in very varied fields. QSAR has a wide range of applications, doesn’t it?
H: Yes, the idea of relating a measure of the behavior of a molecule to its inherent properties is increasingly being used throughout the pharmaceutical, materials and chemical sciences. Non-pharmaceutical applications of this approach are generally referred to as quantitative structure-property relationships, QSPRs, a term I had not heard before I used it. QSPR modeling has been around for quite some time, but has never taken off like QSAR modeling has in the pharmaceutical sciences. But that may be changing. Increasingly chemical companies are turning to QSPR modeling as a way of filtering and selecting trial chemicals in new product development. And of course we see the continued growth of QSPR in the ADME - Toxicology field.

W: As an Associate Editor of the Journal of Chemical Information and Modeling (JCIM), you wrote our new guidelines on QSAR. These have since been taken up by other journals. Tell us briefly what this means to potential JCIM authors.
H: It is extremely important for the continued growth and acceptance of the QSAR/QSPR field that the methods, data and applications reported be validated and be reproducible. My hope, and I think that of my fellow JCIM editors, is that the new QSAR guidelines will provide criteria of standardization to achieve validation and reproducibility of QSAR studies.

W: Why isn’t there a Hopfinger data set?
H: I don’t believe there is a single data set for all applications. For example, to evaluate a method for computing log P would require a different data set than for validating a method to rapidly and effectively do conformational analysis. My strategy has always been to seek out data sets that most meaningfully explore and test a particular theory, method or exploratory idea that we are investigating.

W: Who is your hero, or heroine: the scientist you most admire?
H: I have had the good fortune to work with many excellent scientists over my career. I have tried to let these individuals know my appreciation of, and respect for, what they have done. However, I don’t wish to mention names here.

W: Do you have any retirement plans? What hobbies might you have more time for?
H: I don’t have any retirement plans. My hope is to continue doing what I am doing, but perhaps morphing with time in terms of how much I do and where I do it. In terms of hobbies, I enjoy non-mainstream movies and model trains. Interestingly, my model train hobby came about in part because it was something of an escape from computers and high technology. But now it has become so high tech that doing QSAR research has almost become an escape from high tech model railroading!

W: In parting, do you have any advice for young modelers?
H: I would tell them to take the time to learn enough physical and theoretical chemistry, enough statistics and enough pharmacology so as to be able to understand and critically question the applicability and reliability of the methods being employed in the software they are using or developing. Obviously this is not generally considered fun to do, but I really think modelers today have gotten too far away from a healthy skeptical and questioning attitude toward computational tools. My anchor point is to remember that we still cannot routinely calculate a reliable IC50 value which is the cornerstone measurement in preclinical drug discovery.