Fri, 06 Mar 2026
13:15
13:15
L6
Geometric and topological potentials driving self-assembly
Ivan Spirandelli
(University of Potsdam)
Abstract
The assembly of molecular building blocks into functional complexes is a central theme in biology and materials science. In this talk, we showcase the generative and thermodynamically predictive capabilities of a geometric model, the morphometric approach to solvation free energy, applied to spherical particles, tubes, and protein subunits. We demonstrate that this purely geometric description is sufficient to generate biologically relevant structural motifs and identify native nucleation states in simulation.
However, relying solely on local geometric fit often leads to optimization challenges. Molecular simulations frequently become trapped in local minima because the model lacks global structural information. To address this, we introduce a global bias based on persistent homology. By incorporating a weighted sum of total persistence as an active potential, we obtain an efficient simulation strategy, significantly increasing success rates. Integrating topological descriptions into energy functions offers a general strategy for overcoming kinetic barriers in molecular simulations, with potential applications in drug design, material development, and the study of complex self-assembly processes.