A gene sequence is not the whole story for AI
A gene’s role depends on its cell and on other parts of the genome. Philip Ball explains why models trained on DNA sequences alone cannot capture every link. AI may point researchers toward an answer, but a useful prediction is not the same as knowing the cause or proving a clinical use.
What the source reports
In a science explainer, Philip Ball describes why gene activity is hard to read from DNA sequence alone. A gene’s effect can change with the type and state of a cell. Proteins that help switch genes on and distant stretches of DNA also interact. A model can make a useful prediction from patterns without showing exactly how that chain works. Ball’s argument is an expert explanation, not a head-to-head trial of AI models. It does not establish that a prediction is ready for clinical care.
Original source
- Title
- Why AI Isn’t Ready to Tackle Our Convoluted Genomes
- Author
- Philip Ball
- Publication
- Quanta Magazine / WIRED
- Date
- Saturday, October 17, 2026