The laboratory of bioengineer César de la Fuente is utilizing deep learning models to search for antimicrobial molecule candidates within vast datasets of biological sequences. The laboratory stated that they are using ChatGPT and Codex in tandem for hypothesis generation, the creation and improvement of code, dataset processing, analysis of results, and the synthesis of ideas across different academic disciplines.
Leveraging ChatGPT and Codex for Biological Sequence Analysis to Accelerate Drug Candidate Discovery
This article is a translation. Read the Japanese original
The laboratory consists of members with diverse expertise in biology, chemistry, computer science, and engineering. The use of AI serves to bridge the knowledge gap between members who are proficient in programming but less familiar with biology, and vice versa. Additionally, it is reported that ChatGPT allows researchers to work in their native languages, contributing to an accelerated workflow.
On the other hand, de la Fuente cautioned against relying solely on AI, stating that the accuracy of predictions must be constantly re-verified. He pointed out that empirical experiments (ground truth) are essential to validate AI predictions.
Sources
- How a researcher uses Codex and ChatGPT to search for new antimicrobial molecules （OpenAI News, 2026-09-10）