Ligand-based drug design in the AlphaFold age

Hello, my name is David Clark. I work on computer-aided drug design for Charles River Laboratories. I'm going to talk now about the topic of ligand-based drug design in the AlphaFold age.

The talk breaks down into three sections: A very short introduction that I've called Meet Cinderella, the main part of the talk is the section called Reasons to be Cheerful, and then there is just a single-slide conclusion at the end.

Let's set the background. This is what I've called Meet Cinderella. I think in some ways, you've really got to feel sorry for ligand-based drug design at the moment. There are so many aspects of structure-based drug design that have come to the fore over the recent years that it's very easy to forget and overlook ligand-based drug design. That's why I've thought of ligand-based drug design as a bit of a Cinderella figure. You can see a picture of someone like Cinderella in the middle of the slide there, feeling rather sorry for herself, and that's because of all these things around the outside of the slide. For instance, the protein databank of publicly available X-ray crystal structures of proteins has grown year on year remarkably until it passed over 200,000 structures back in January 2023. As if that wasn't enough, you cannot fail to have heard of the AlphaFold program: AlphaFold2, and most recently AlphaFold3. With AlphaFold2, Google's DeepMind was able to predict more than 200 million three-dimensional structures for proteins, giving a huge boost to structure-based drug design. Then added to that, a new experimental technique called cryo-EM has come to the fore, and that's working really nicely in complement to X-ray crystallography to provide atomic resolution models of previously very difficult study proteins. Finally, from a computational aspect, the application of free energy perturbation has, at least in some instances, shown potential for accurately predicting the binding affinity of compounds in structure-based drug design projects. All these things combine really to boost the profile and the power of structure-based drug design, and as a result, it's very easy to forget about or to overlook ligand-based drug design methods. But as I hope to show today, and as perhaps the clever rat in the picture says, well, maybe it's not all over for ligand-based drug design. Let's look at the reasons why we should not forget ligand-based drug design and continue to use it in our projects.