Machine Learning for Medicine: Accelerating Metadynamics of Supramolecular Host-Guest Complexes for Therapy and Imaging

In the context of 3-dimensional porous supramolecular materials (metallacages) computational methods hold promise to accelerate the challenging process of designing and optimizing their host-guest chemistry and encapsulation properties. Amongst the possible atomistic simulation approaches, metadynamics (metaD) has proven very powerful. However, the choice of high level chemical descriptors, i.e. the collective variables (CVs), to accelerate the numerical computation is not trivial for complex systems. To automate the process of CVs identification, we apply here different machine learning algorithms to the metaD simulation of metallacages designed for biomedical applications, including drug delivery, imaging and therapy.