Machine learning integrated with in vitro experiments for study of drug release from PLGA nanoparticles.

This paper investigates delivery of encapsulated drug from poly lactic-co-glycolic micro-/nano-particles. Experimental data collected from about 50 papers are analyzed by machine learning algorithms including linear regression, principal component analysis, Gaussian process regression, and artificial neural networks. The focus is to understand the effect of drug solubility, drug molecular weight, particle size, and pH-value of the release matrix/environment on drug release profiles. The results obtained from machine learning is then used as guidelines for designing new in vitro experiments to examine dependence of drug release profiles on those four factors. It is interesting to see that indeed the results of the new in vitro experiments are in basic agreement with the results obtained from machine learning.