How a 17-Year-Old Turned Proteins Into Numbers and Won $90,000

 

Jerry Xu

What If You Could Turn a Complex Protein Into a Simple String of Numbers?

Imagine taking one of nature's most complex 3D puzzles, a protein structure, and translating it into a clean string of numbers that a computer can instantly read and compare. That is exactly what 17-year-old Jerry Xu from Lexington, Massachusetts did. His breakthrough AI model just earned him $90,000 at the prestigious Regeneron Science Talent Search.

The Real Problem Slowing Down Biotech Discovery

In computational biology, structure equals function. If you know how a protein is shaped, you can better understand what it does. The problem is scale. Modern biology generates massive databases of molecular structures, and current tools are stuck between two bad options: they are either too slow to handle the volume, or they are fast but miss critical structural details. This bottleneck slows everything from basic research to drug discovery.

Advert

Inside the Breakthrough: A Unified Protein Embedding Model

Xu's project, titled "A Unified Protein Embedding Model With Local and Global Structural Sensitivity," solves this with elegant efficiency. 

His AI model learns to distill both local and global structural information into unified numerical representations, or embeddings. Instead of re-analyzing every fold and atom from scratch, researchers can now use these number strings to rapidly find similarities between proteins. It is particularly powerful for studying mysterious protein regions with unknown functions by comparing them to well-studied molecules and getting instant clues.

Meet Jerry Xu: Scientist, Educator, and Civic Coder

Jerry is a senior at Lexington High School and captain of both the math team and science bowl club. But he is not just focused on the lab.

He is passionate about making science accessible, co-founding a summer science lecture series for elementary to high school students and working as a teaching assistant at KTBYTE Computer Science Academy, where he teaches biology, math, and live experimental science.

And his coding goes beyond biology. For the past two years, he has led a team of student programmers building a free AI chatbot that helps local residents understand their town's municipal budget. From molecules to municipal finance, he is using AI to solve real problems.

Why This Innovation Matters More Than Ever

Jerry Xu's work is a perfect example of the future of AI in life sciences. By turning complex biology into efficient, searchable data, models like his could dramatically accelerate how we compare proteins, predict functions, and discover new therapeutics.

It also proves that big ideas do not require big labs. Sometimes, they just need a fresh perspective, a clever algorithm, and a 17-year-old who knows how to think outside the box.

What do you think is the next field AI will transform?

#ArtificialIntelligence #Biotech #ComputationalBiology #ProteinFolding #DrugDiscovery #RegeneronSTS #STEM #AIInnovation #Bioinformatics #FutureOfScience #JerryXu