Extended Data Table 1 Drug-likeness related and target-centric performance of DrugGEN and other methods: RELATION34, ResGen28, TRIOMPHE-BOA92, TargetDiff29, and Pocket2Mol15, measured in terms of of QED, synthetic accessibility (SA), FCD, fragment similarity, scaffold similarity, and adherence to Lipinski, Veber, and PAINS filters, together with docking scores (median kcal/mol values of the top 10% molecules in terms of docking scores), for the AKT1 and CDK2 targeting tasks, separately

From: Target-specific de novo design of drug candidate molecules with graph-transformer-based generative adversarial networks

  1. Values for the training dataset (real molecules) are also provided. The arrows indicate the direction in which the performance increases. The best performances are displayed in bold font, separately for AKT1 and CDK2 targeting models, where all scores within 1% of the highest value are included, as the differences among these close values are negligible. Performance values that could not be calculated are indicated by ‘-’. Metrics that contain a distribution of values are reported as mean ± SD, while metrics yielding a single value are reported as such.
Back to article page