Rational Design of Breast Cancer-Selective Lipid Nanoparticles via Multi-Objective Bayesian Optimization

We have previously demonstrated that single-objective Bayesian optimization is an effective framework for rational optimization of lipid composition of LNP to enhance delivery to target cells.^[1] However, single-objective Bayesian optimization was insufficient to suppress delivery to non-target cells. To approach this problem, we employed multi-objective Bayesian optimization, which simultaneously maximizes delivery to target cancer cells and minimizes uptake in normal non-target cells.