Dissertation
Computational tools and protocols for the development of photoactivated metallodrugs
Metal-organic compounds hold great promise as photoactivated chemotherapy (PACT) agents, leveraging their unique geometric and electronic properties to achieve precise spatiotemporal control over drug activation.
- Author
- M.L.A. Hakkennes
- Date
- 17 April 2026
- Links
- Thesis in Leiden Repository
However, their computational design has been hampered by the lack of dedicated screening tools, limited biological activity data, and the complexity of modeling photosubstitution reactions in realistic environments.This thesis addresses these challenges through three interconnected approaches. First, MetalDock was developed, a docking tool tailored for metal-containing drugs that derives missing Lennard-Jones parameters for twelve metal atom types via Monte Carlo sampling, achieving high spatial accuracy across diverse biomacromolecular targets. Second, transfer learning was shown to effectively bridge organic and metal-organic chemical spaces for predicting protein inhibition (pIC50/pKi), establishing domain adaptation as a viable strategy for data-scarce metallocompound modeling. Third, two complementary protocols were developed to elucidate photosubstitution mechanisms: a static DFT approach incorporating explicit solvent that linked computed barriers to experimental thermal barriers, and AQUAMEPTH, a semi-empirical molecular dynamics protocol that revealed bidentate ligand planarity, rather than the conventional ³MLCT–³MC barrier, as the key determinant of photosubstitution efficiency.Together, these tools provide a computational framework spanning docking, activity prediction, and mechanistic insight for next-generation PACT agent design.