FORGE
Binding tells you something. But it does not predict whether a therapy will activate T cells, drive killing, or hold up as cancer changes. Coding Bio built FORGE to measure that directly — in living immune cells, at a scale the field has not had before.
FORGE evaluates over one million candidates directly in primary human immune cells — measuring T-cell killing, activation, and cytokine profiles across more than 50 million datapoints. Not whether a molecule binds to a cancer cell. Whether it drives immune response.
Our machine learning models are trained on these functional readouts, learning which molecular designs are associated with immune response. They use that to guide the architecture of the next generation of candidates — improving with every screening cycle.
Targeting cancer from multiple angles requires complex, multi-modular molecules — and precise control over how they act. OR-gate logic preserves activity across more than one disease-associated signal, making escape harder as cancer adapts. AND-gate logic adds a further layer of precision: ensuring healthy cells are not caught in the crossfire.
This is what it means to make the immune system programmable.
The result: a pipeline built on immune response, not binding alone.
See the Pipeline →