10.1101/2023.10.03.560711

A cautionary tale about properly vetting datasets used in supervised learning predicting metabolic pathway involvement