AI Helps Lab Find New Antibiotics in Hours, Not Years

Drug-resistant infections killed five million people in 2021, and that number could double by 2050. Yet we haven't discovered a new class of antibiotics in 50 years.

César de la Fuente, a bioengineer, believes he's found a faster way forward. His lab uses artificial intelligence to search through massive genome databases for molecules that could become tomorrow's lifesaving medicines.

The traditional approach to finding new antimicrobials can take years of painstaking work. Scientists collect samples from plants, animals, and soil, then test candidate molecules one by one in a slow, iterative process.

De la Fuente's team takes a different path. They treat biology as an information system, where DNA nucleotides and protein amino acids function like an alphabet.

Their deep-learning models scan vast datasets of genetic code, searching for patterns that signal a molecule might fight infections. What once took years now happens in hours.

The lab uses ChatGPT and Codex alongside their custom AI models to brainstorm ideas, write code, process datasets, and connect insights across different scientific fields. Team members with biology backgrounds use AI to help them program, while computer scientists lean on it to understand biochemistry.

De la Fuente uses ChatGPT as a brainstorming partner, feeding it ideas to help shape hypotheses. His diverse team contributes their thoughts to a shared ChatGPT workspace, creating a collaborative sounding board that bridges different areas of expertise.

But AI is just the starting point. Once the models identify promising candidates, lab scientists must confirm the molecules actually kill microbes and test their effects on human cells.

Successful candidates then face years of additional testing for toxicity, manufacturing viability, and regulatory approval before reaching patients. That's why de la Fuente insists that AI and traditional lab work must advance together.

The Ripple Effect

The lab's approach doesn't just speed up drug discovery. It opens entirely new spaces for exploration by searching the genomes of both living and extinct organisms.

De la Fuente sees this work as scratching the surface of understanding biology, which he calls "the most complex thing out there." His transdisciplinary team proves that AI can lower barriers between scientific fields, letting experts collaborate in ways that weren't possible before.

Lab members even work in their native languages with ChatGPT, removing another obstacle to groundbreaking research. And while de la Fuente cautions that everyone must double-check AI for accuracy, he values how it democratizes access to the world's scientific knowledge.

The race against antimicrobial resistance just got faster, and the tools to win it are finally catching up to the challenge.