Emeryville, California-based Profluent and BioMarin Pharmaceutical (Nasdaq: BMRN) entered a research collaboration under which Profluent will apply its AI protein language models to design and optimize novel enzyme therapy candidates for BioMarin's pipeline, with financial terms not disclosed.
Profluent's platform centers on proprietary protein language models trained on the Profluent Protein Atlas, which the company said contains more than 115 billion unique protein sequences. The models will generate de novo enzyme candidates targeting desired activity and stability profiles. BioMarin will screen those outputs and advance the most promising designs into preclinical testing. No specific disease indication, gene target, or named program has been disclosed, with the collaboration set to target genetically defined conditions.
BioMarin's existing enzyme replacement therapy portfolio spans several rare inherited metabolic disorders, and the collaboration is intended to expand the range of enzyme candidates accessible through AI-guided protein design.