"Designing Life with AI" is a cross-disciplinary MAKE project in which EPFL students get a chance to explore the research process by designing proteins. A total of eight research labs are involved.
Our body has some 20,000 different kinds of proteins: collagen, insulin, hemoglobin and many more. These molecules perform the myriad functions that are essential for our cells to survive. Thanks to advances in artificial intelligence, scientists are gaining valuable insight into the world of proteins and their various structures. Each protein has its own 3D structure that is associated with its function. Today, scientists can model these structures based on the amino acids sequences, and they can create new proteins designed for specific functions. All this opens up promising avenues of research in the field of protein design. This is the research opportunity that Sahand Jamal Rahi, an assistant professor and the head of EPFL's Laboratory of the Physics of Biological Systems, wanted to give EPFL students through the project "Designing Life with AI".
Get hands-on experience in research
"When I saw that AI-driven software was available for designing proteins in a fairly easy way, I realized I could combine my interest in the topic with my goal of giving more students experience with the research process," says Rahi. "I've always encouraged students - especially those in my first-year thermodynamics class - to join EPFL research labs. I think it's extremely valuable for young people to get hands-on experience in research, since reading about research is one thing, but applying it in practice is quite another."
"Designing Life with AI", which is supported by EPFL's MAKE initiative, gives students the opportunity to carry out research projects involving protein design. Eight research labs and around 30 people are currently involved. Cris Darbellay and Mateo Schärer Gonzalez, bachelor's students in life science engineering, are the project leaders for "Designing Life with AI". They explain: "Beyond the research itself, students also get a chance to meet with professors and network informally. We also provide procedures for the lab experiments and user guides for the software."
When I saw that AI-driven software was available for designing proteins in a fairly easy way, I realized I could combine my interest in the topic with my goal of giving more students experience with the research process.
Creating proteins for specific purposes
The project that Gonzalez and Darbellay are working on relates to signaling proteins such as kinases, which play a key role in regulating cell function. They're studying how light-oxygen-voltage-sensing domains (LOV domains) can be added to these proteins in order to regulate their activity. When LOV domains are exposed to blue light, for example, they change shape and can alter the protein's state. "We'd ideally like to create a kinase that contains a LOV domain, so that the kinase can be activated by applying blue light," says Darbellay.
Other "Designing Life with AI" projects relate to binders, compounds that can bind to specific proteins. Scientists use binders to identify a particular toxin, for example, block a protein or change a protein's signaling pathways. Alexia Möller, a master's student in life science engineering, and Dario Sergo, a master's student in physics, are working on nanobodies, which are small fragments of antibodies. The two students are developing a fluorescent nanobody that uses a self-penetrating peptide to enter into cells. Their mechanism would enable scientists to observe protein interactions within human cells. "Our goal is to create a method for designing nanobodies that are tailored to individual antigens," says Möller.
Proteins are central to the biological processes that support life. Breakthroughs in the study of proteins have applications in an array of areas, from disease detection and treatment to environmental remediation and carbon capture. At EPFL, a growing number of students are taking an interest in the vast potential of these fascinating molecules.
