Machine learning approaches to quantitively predict selectivity of compounds against hDAC1 and hDAC6 isoforms

Abstract

AbstractThe design of compounds selectively binding to specific isoforms of histone deacetylases (hDAC) is an ongoing research to prevent adverse side effects. Two of the most studied isoforms are hDAC1 and hDAC6 that are important targets to inhibit in various disease conditions. Here, various machine learning approaches were tested with the aim of developing models to predict the bioactivity and selectivity towards specific isoforms. Selectivity models were developed by directly training on the bioactivity differences of tested compounds against hDAC1 and hDAC6. Both classification and regression models were developed and compared to each other by using traditional evaluation metrics.