Synthyra Information

Synthyra is a Public Benefit LLC building protein design technology that optimizes whole biological systems. In the limit, that means steering an entire metabolism toward any outcome you choose: - A cancer therapeutic that minimizes off-target toxicity - A therapeutic that takes flux through a viral polymerase to zero - An algal metabolism tuned for carbon capture and biofuel at minimum cell burden Binder design works. Enzyme design is arriving. High-fidelity control over almost any protein property is inevitable, and it leaves the harder question untouched: do these campaigns solve the problem they were started for? A good binder and a good biological outcome are not the same thing. If the goal is a disease response, make that the objective. That takes biological context at a scale nobody could afford before, so we built it. Our protein-protein and protein-ligand models score billions of interactions per second at 13% higher performance than the best published methods. A whole interactome becomes an object you optimize over, not a dataset you curate. Our annotation models predict most protein properties from sequence alone. We are now integrating genome-scale metabolic models into the same objective. Decades of curation made them our most rigorously grounded account of cellular metabolism, bounded by what has been annotated and measured. Our preliminary work fits omics data across chemically plausible reaction space, extending rather than replacing them. We learn continuous representations of how proteins associate with abstract labels like Gene Ontology and disease, making that objective computable. Early pipelines have already surfaced verified mechanisms behind human disease and symptoms, and produced picomolar binders with better systems profiles. We are raising to build the rest. Run it yourself at synthyra.com, via an agent-friendly API or the UI. Or run a program with us: therapeutic, enzyme, metabolic.