Effective Holistic Characterization of Small Molecule Effects Using Heterogeneous Biological Networks

The study utilizes the CANDO platform to enhance drug discovery by integrating multiscale interactomic signatures from heterogeneous biological networks, analyzing interactions between 12,951 drugs/compounds and the human proteome. This approach improves drug repurposing accuracy and adverse drug reaction (ADR) prediction, achieving a mean top10 average indication accuracy of 30.0%. The multiscale method outperforms traditional proteomic and chemical pipelines, particularly for complex indications like neoplasms and mental disorders, though proteomic signatures perform better for infectious diseases. The study highlights the potential of interactomic networks to provide detailed mechanistic insights and improve drug repurposing accuracy, while acknowledging limitations such as the need for more comprehensive data. Future work aims to enhance these networks with additional data and advanced modeling techniques.