The report positions IsoDDE as extending work begun with AlphaFold 3 (AF3). While AF3 advanced protein structure prediction, Isomorphic Labs argues that accurate structures alone are insufficient for many prospective in silico drug-discovery workflows; IsoDDE is presented as an integrated engine for ligand–target modelling and design.
According to the document, IsoDDE is intended to predict binding-related properties and to support exploration of protein sites, including so-called cryptic pockets, in unified workflows. The authors compare performance and throughput to selected physics-based and machine-learning baselines described in the report.
The company outlines ongoing internal programmes and states that some AI-assisted drug candidates are progressing toward clinical evaluation; specific timelines and indications are described in the technical report rather than in this summary.
For architecture, benchmarks, limitations, and stated use cases, see the IsoDDE technical report (PDF) (Isomorphic Labs).