De novo nanobody design generating epitope-targeted VHH binders from a target sequence, with only 14-50 candidates sent for experimental testing.
All-atom antibody-antigen complex structure prediction on an AlphaFold 3-inspired architecture, served as a closed model through MoleculeOS.
Reactive machine-learning force field running DFT-accurate molecular dynamics in which bonds break and form, from proton hopping to enzyme catalysis.
Multimodal protein foundation model spanning sequence, structure, function, and evolution, used for de novo design and structure prediction.