Ethan Tran stands in a library. He smiles at the camera.

The Math of Molecules

Tran Helps Drug Discovery Startups

By Kirsten Heuring Email Kirsten Heuring

Heidi Opdyke

Carnegie Mellon University undergraduate Ethan Tran is using mathematics to help chemistry startups discover new reactions, advanced materials and potential drug candidates.

Tran, a rising junior in mathematical sciences, works with startups: Cavall Labs and Rhizome Research. Besides developing AI-driven tools for researchers, he helps scientists navigate complex chemical data more efficiently and find promising discoveries faster.

“Some researchers aren’t fully accepting of computational tools since some of the time, the tools are slow and inaccurate,” Tran said. “There’s a lot of work to be done there, and I want to apply computational work to the hard sciences.”

Tran was a founding contributor of Cavall Labs, a startup created by Department of Chemistry Ph.D. student Rajdeep “Ron” Sarma with support from CMU VentureBridge, a Swartz Center for Entrepreneurship program that supports high-potential ventures founded by Carnegie Mellon graduate students and professors. Cavall Labs develops machine learning technologies to accelerate molecular discovery by helping chemists identify promising molecules faster. Their tools can predict if reactions are likely to occur, identify reactive regions within molecules and curate data. Each of these tasks conventionally took days to weeks, but Cavall can do them in seconds to minutes.

Tran’s work started before Cavall was formed, when he was conducting research with Isaac Garcia-Bosch, associate professor of chemistry and a faculty advisor to Cavall Labs. Working with Garcia-Bosch, Tran created a program foundational to the startup’s first product.

Project Effective QUest into boUnded chemical Spaces (EQUUS), an open-source workflow, screens molecules to determine their potential as redox mediators in chemical reactions. Tran helped develop the Adaptive Lightweight Optimization Engine (ALOE), which is the backbone of Project EQUUS, ranks potential redox mediators at ultrafast speeds with quantum accuracy. He also created a system that is highly modular and automatically adjusts the program’s computing power requirements based on the input size and available hardware, making the tool highly effective and accessible to chemists with varying levels of computational resources.

Using the program, the Cavall team identified molecules that could be used in next-generation batteries. Sarma said that Tran’s perspective has been instrumental in advancing Cavall’s work.

“Ethan’s mathematical thinking pushes not only me but our entire team to come up with distinctively alternate thinking which has so far helped us immensely,” Sarma said. “Ethan is well beyond his age in terms of maturity of thought, his hardworking nature and open-minded curiosity in applying rigorous mathematical reasoning in anything he does.”

Tran’s work with Cavall Labs led him to a second chemistry startup: Rhizome Research, which uses artificial intelligence tools to accelerate drug discovery. Yiwen Wang, a Department of Chemistry Ph.D. alumnus as well as the cofounder and chief scientist of Rhizome Research, was a teaching assistant in a chemistry course that Tran took. After working closely with Tran, Wang invited him to join Rhizome Research.

At Rhizome, Tran evaluates emerging AI technologies and adapts them to improve the startup’s research tools and workflows.

“Technology is advancing so rapidly, and I'm architecting our systems to keep up at scale,” Tran said. “We leverage the best of cutting-edge methods and tools.”

Rhizome’s tools help researchers find molecules that they can create in the lab, accelerating a process that typically takes years to only take months or weeks. The software automatically identifies similar molecules that may bind effectively to cells’ receptors, allowing researchers time to focus on experiments or other research priorities. Tran developed software that allows many of these processes these tools to run autonomously and deliver relevant results to researchers.

“These models couldn’t differentiate between the different orientations of a molecule,” Tran said. “Inherently, there’s a math problem there. A molecule’s structure is like a graph. I figured out a way to fragment a molecule so the models can process it.”

David Offner, teaching professor and director of undergraduate studies for the Department of Mathematical Sciences, said Tran’s work demonstrates how mathematical training can advance research across disciplines.

“Expertise in mathematics, and the reasoning and quantitative skills developed by studying math, can be applied in almost any domain,” Offner said. “Ethan’s work is helping small, interdisciplinary companies that are moving quickly to take advantage of new opportunities created by artificial intelligence.”

Tran plans on working with both Cavall and Rhizome Research. He said that Carnegie Mellon’s collaborative culture for creating opportunities prepared him for interdisciplinary research now and in the future.

“I chose CMU because I wanted the interdisciplinary environment,” Tran said. “It’s an environment that encourages people from different backgrounds to come together. I got very lucky because that’s exactly what I’m dealing with.”