Interactive Knowledge-Based Kernel PCA for Solvent Selection

Abstract

Selecting more sustainable solvents is a crucial component to mitigating the environmental impacts of chemical processes. Numerous tools have been developed to address this problem within the pharmaceutical industry, employing data-driven approaches such as multidimensional scaling or principal component analysis (PCA). Interactive knowledge-based kernel PCA is a variant of PCA that allows users to shape 2D solvent maps by defining the positions of data points, imparting expert knowledge that was not included in the original descriptor set. We have applied interactive PCA to the task of solvent selection and present an intuitive interface that is integrated into AI4Green, an electronic laboratory notebook that encourages sustainable chemistry. A set of evidence-based user guidelines were developed and used in combination with the interactive PCA to identify four potential solvent substitutions for an example thioesterification reaction.