Diffusion Models at the Drug Discovery Frontier: A Review on Generating Small Molecules Versus Therapeutic Peptides

Diffusion models have emerged as a leading framework in generative modeling, holding significant promise to transform the traditionally slow and costly process of drug discovery. This review systematically compares their application in designing two primary therapeutic modalities: small molecules and therapeutic peptides. The unified framework of iterative denoising is adapted to the distinct molecular representations, chemical spaces, and design objectives inherent to each modality. For small molecules, these models demonstrate proficiency in structure-based design, facilitating the generation of novel, pocket-fitting ligands with desired physicochemical properties, though ensuring chemical synthesizability remains a critical challenge. Conversely, for therapeutic peptides, the emphasis shifts towards generating functional sequences and de novo structures, where primary

challenges include achieving biological stability against proteolysis, ensuring proper folding, and minimizing immunogenicity. Despite these modality-specific distinctions, both domains confront shared hurdles, such as the scarcity of high-quality experimental data, reliance on potentially inaccurate scoring functions for validation, and the imperative for rigorous experimental validation. Unlocking the full potential of diffusion models necessitates bridging these modality-specific gaps and integrating them into automated, closed-loop Design-Build-Test-Learn (DBTL) cycles.

Message: Hi FE, I need help with “Diffusion Models at the Drug Discovery Frontier: A Review on Generating Small Molecules Versus Therapeutic Peptides” in “Biotechnology & Biomedical Engineering”