Profluent announced o ProGen3, a new family of artificial intelligence (AI) capable of designing complex proteins from scratch – with the results marking the first evidence of AI scaling laws in biology, proving that larger models and more data create more robust results.
ADVERTISING
ProGen3 Details
- The biotech company's 46-billion-parameter model was trained on 3,4 billion protein sequences, outperforming previous datasets and demonstrating improved protein generation.
- He successfully designed new antibodies that performed equivalently to approved therapeutics, but were distinct enough to avoid patent conflicts.
- The platform has also created gene-editing proteins less than half the size of CRISPR-Cas9, potentially enabling new delivery methods for gene therapy.
- Profluent is making 20 “OpenAntibodies” available through royalty-free or upfront licensing, targeting diseases affecting 7 million patients.
Why is it important
If scaling trends hold, Profluent’s approach could transform drug design and gene editing from years of lab work into a simple engineering problem.aria faster and more predictable – rewriting how new therapies are discovered. These trends also suggest that we are just at the beginning of AI’s impact on drug discovery and medicine.
Read also
