TRION Cloud

Browser-based platform for computer-aided drug discovery that runs structure-based and AI workflows without local installation. AI cofolding predicts the 3D structure of a protein complex with a ligand, DNA or RNA using Boltz-2, together with small-molecule binding affinity. Molecular docking with GNINA screens a compound library against a protein pocket and ranks hits by score, drug-likeness and chemical family. Target-specific activity prediction fine-tunes the Boltz-2 affinity predictor on the user's measurements or builds a docking-based QSAR model with interpretable interaction and shape features, and pre-built models are available for targets such as EGFR, MDM2 and LRRK2. Compound generation proposes new molecules with REINVENT4, compound combination enumerates R-group combinations of two compound series, compound alignment places compounds onto a reference core, and compound design builds analogs from a reference ligand. Results can be viewed in 3D and downloaded.