[agsandrew/Adobe Stock]
“You’re in a dark room, fumbling around for the light switch,” said Paras, a professor at the Institute for Neurodegenerative Diseases at University of California, San Francisco (UCSF). He suggests that the early stages of drug discovery can, in some ways, mirror the experience of navigating the dark, puzzling dungeons of a game like The Legend of Zelda, where players confront a complex, unknown environment. Much like this, researchers often find themselves in the vast unknown of cellular assays and biochemical screens, seeking potential therapies.
How SBDD transformed drug discovery
But the advent of structure-based drug design helped illuminate the field, turning drug discovery from a quest in darkness to more of a guided journey. With structure-based drug design (SBDD), it feels like turning on the light in a dark room, Paras said. “Once we got to structure-based drug design, it’s like oh, there’s a pocket. There’s space over here,” he said. “There’s a hydrogen bond donor over here.” SBDD, a computational approach used in drug discovery to design molecules that can interact with specific biological targets, turned on the metaphorical light switch. “You turn the light on, and it’s a completely different way to approach the problem,” Paras said.
SBDD relies on knowledge of the 3D structure of the target obtained from methods like X-ray crystallography or cryo-electron microscopy. “X-ray crystallography was the first way that we could see atomic level resolution structures of biomolecules,” Paras said. “That was a revolution in the field.” Thanks to SBDD, researchers can design molecules based on an understanding of the target’s structure. That is, they can ensure a precise fit into the active site or other relevant regions, which can then modulate the target’s function. The technique helped accelerate drug discovery by reducing the need for time-consuming trial and error methods, allowing for a more efficient and effective approach.
Advanced technologies like molecular dynamics (MD) simulations have also had a significant impact on drug development. These computer simulations are used to study the physical movements and interactions of atoms and molecules over time, providing vital insights into the stability and effectiveness of potential drug molecules with their target proteins.
Two physicists, Berni Alder and Tom Wainwright, proposed the concept of molecular dynamics in the 1950s. The technique would go on to have a significant impact on drug discovery.
Cryo-EM and the molecular world
This composite image of beta-galactosidase highlights precision of cryo-EM’s resolution. [Image Courtesy: Veronica Falconieri, from the Subramaniam Lab at the National Cancer Institute.]
“There’s something really exciting about the revolution in bio-electron microscopy, particularly for neurodegeneration research. Being able to see atomic-level resolution structures at 2.7 to 2.5 angstrom resolution of these highly ordered biological polymers is truly amazing,” Paras said.
Cryo-EM has been particularly valuable in studying large protein complexes and membrane proteins that were previously difficult to study using traditional X-ray crystallography. This revolutionary imaging technology has enabled researchers to gain deeper insights into the molecular mechanisms underlying various biological processes, thus supporting the development of new and improved therapeutics.
“I would never have imagined that a microscope could show me atoms,” Paras said.
“There’s something really exciting about the revolution in bio-electron microscopy, particularly for neurodegeneration research. Being able to see atomic-level resolution structures, around 2.7 to 2.5 angstrom resolution, of these highly ordered biological polymers is truly amazing. As a young chemist getting into this field, I would never have imagined that a microscope could show me atoms.”
Enter Sandbox AQ: AI meets quantum-inspired drug discovery
In an ongoing search to further their research, UCSF’s Institute for Neurodegenerative Diseases explored numerous avenues. Paras revealed, “In the lab, you know, we’re starting to think about things we could do with biophysical tools and other screening methods to get to it, but nothing was really working well.”
That’s when they discovered Sandbox AQ. “It was just kind of like a shot in the dark,” Paras confessed. Nonetheless, the risk paid off. UCSF’s Institute began leveraging Sandbox AQ’s AI-powered quantum simulation technologies, including their proprietary Absolute Free Energy Perturbation (AQ-FEP) software, and saw immediate improvements.
This platform, which applies the principles of SBDD and predicts quantum-mechanical interactions between atoms and molecules, offered a boost to their drug discovery process. Paras noted, “We found that, really working with Sandbox was the first time we could make the assumptions match reality with what we were saying, and that allowed us to really start to do experiments in-silico that were meaningful.”
The real-time response of the software enabled the team to prioritize their best ideas and improve the quality of their outcomes. “If each round of iteration you’re taking, not just 10 ideas, but your top 10 out of 100 ideas, you’re putting yourself in a better position to succeed,” Paras said. “You’re going to get higher quality hits out of your screening collection, faster improvements in potency, and you’re going to be able to prioritize compounds that can move faster in the clinic.”
Paras emphasizes the urgent need for further advances in the field. “We need to be looking for any advantage, any technology that can help us get to an actionable hypothesis and a drug that can be tested in the clinic as quickly as we can. Otherwise, this will never be solved in our lifetime, our children’s lifetime, or their children’s lifetime.”
Filed Under: Data science, machine learning and AI, Neurological Disease
