Daniel Reker: Using Machine Learning to Explore New and Existing Therapies

May 29, 2020

New faculty member Daniel Reker uses active machine learning to analyze and design new drug therapies for use against cancer and infectious diseases

Daniel Reker

Daniel Reker

Daniel Reker will join Duke University’s Department of Biomedical Engineering beginning September 1, 2020. With research at the intersection of computational and molecular biology, data science and engineering, Reker specializes in using machine learning and modeling to explore how drugs, excipients and nanoparticles behave once they enter the body.  Reker will contribute his expertise to a growing community of faculty focused on biomedical and health data science at Duke.

Daniel Reker

Assistant Professor of Biomedical Engineering (September 1, 2020)

Hometown: Frankfurt, Germany

Alma Maters: TU Darmstadt, ETH Zurich

Representative Publication: Reker et al. bioRxiv 2019, https://doi.org/10.1101/786251.

Fun Facts: Marathon runner (3:38); won 12 medals as member of the Hessian National competitive ballroom dancing squad

Currently a SNSF postdoctoral fellow at the Massachusetts Institute of Technology, Reker works in the lab of Robert Langer—the most cited engineer in history—at the Koch Institute for Integrative Cancer Research. Reker’s primary projects have involved developing new machine learning approaches to design novel formulations and self-assembling nanoparticles that can more effectively deliver drugs. He uses a similar approach to explore the biological effects and limitations of therapies for different types of cancer and infectious diseases like malaria, HIV and­­­­—most recently—COVID-19.

Reker will continue this work in his new role at Duke, where his lab will combine computational work with wet-lab experiments to analyze and design drug therapies.  A key component of this research involves active machine learning. Traditional machine learning algorithms work with large data sets to develop formulas and make predictions. But with active machine learning, the algorithms are able to ask questions or request that the researchers provide more information if the algorithm has identified a gap in the data. This allows the model to be more efficient, and allows for more comprehensive prediction.

“There are still practical concerns that need to be addressed with this approach, but my lab will mature it and apply it to create models for how molecules move through and interact with the body,” says Reker. “For example, we could study a medication that failed in a clinical trial and use active machine learning to explore specifically why the drug failed, whether it was a negative drug action or an issue with the body’s biology. That knowledge could help us salvage the failed drug by tweaking it or by generating novel delivery modes to create a more effective therapy.” 

Why Duke?

“Duke was a huge draw for me because it’s such a uniquely collaborative environment, where I can easily work with researchers across different departments or simply walk across the street if I want to collaborate with any clinical partners at the hospital.”

—Daniel Reker

Although a majority of his current work takes place in the lab, he’s also fostered collaborations with clinicians and pharmacists at the Brigham and Women’s Hospital in Boston to better understand how his techniques could translate to patients. Now, Reker is looking forward to bringing a similar collaborative spirit to Duke.

“My work involves chemistry, bioinformatics, pharmacology, computer science and biomedical engineering, and that means that collaboration is key,” he says. “Duke was a huge draw for me because it’s such a uniquely collaborative environment, where I can easily work with researchers across different departments or simply walk across the street if I want to collaborate with any clinical partners at the hospital.

“I’m also excited to be working so close to Research Triangle Park, which will hopefully enable lots of cool interactions and make potential collaborations easier. A lot of my work is relevant to the pharmaceutical and biotech industry, so this is really a prime place for me to be.”

For Reker, who received both his PhD in Pharmaceutical Sciences and a master’s degree in computational biology and bioinformatics from ETH Zurich, his new role also provides him with an exciting opportunity to mentor and connect with students.

“I supervise students with backgrounds ranging from experimental biology and chemistry to computer science. Working with students and being able to help guide their career paths was something that always attracted me to academia,” he says. “My undergraduate degree was in computer science, and I’ve been able to coalesce different fields through my academic career because of the support I had. Now, I want to make sure that students feel empowered and supported to choose their own path.”