Feasibility of Active Machine Learning for Multiclass Compound Classification

Abstract

A common task in the hit-to-lead process is classifying sets of compounds into multiple, usually structural classes, which build the groundwork for subsequent SAR studies. Machine learning techniques can be used to automate this process by learning classification models from training compounds of each class. Gathering class information for compounds can be cost-intensive as the required data needs to be provided by human experts or experiments. This paper studies whether active machine learning can be used to reduce the required number of training compounds. Active learning is a machine learning method which processes class label data in an iterative fashion. It has gained much attention in a broad range of application areas. In this paper, an active learning method for multiclass compound classification is proposed. This method selects informative training compounds so as to optimally support the learning progress. The combination with human feedback leads to a semiautomated interactive multiclass classification procedure. This method was investigated empirically on 15 compound classification tasks containing 86-2870 compounds in 3-38 classes. The empirical results show that active learning can solve these classification tasks using 10-80% of the data which would be necessary for standard learning techniques.

Active Learning for Computational Chemogenomics

“…Rarey and coworkers recently published an active learning framework for a multitarget problem in drug discovery, namely the ability of active learning combined with chemogenomic reasoning to build a predictive model for a target subfamily by using only a subset of available data [56]. Their investigation demonstrated efficient navigation of focused ligand-target spaces and provided a means for model training and experimental design.…”

“…That is, we investigate the potential of active learning to act as the steering wheel for family-wide computational chemogenomic model building. We evaluate its ability to predict bioactivity and how it evolves a chemogenomic model [49,[56][57][58], finding that protein/target family-wide chemogenomic active learning can build an interaction model from only a small fraction of bioactivity data points in a screening database. These models show high predictive performance on datasets many folds larger than that used for model construction.…”

Small Random Forest Models for Effective Chemogenomic Active Learning

“…To date, several pro-and retrospective studies report successful applications of AL strategies throughout different fields of research [27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42][43][44][45][46]. In the area of drug discovery, active learning has been shown to efficiently derive high-performance prediction models based on small subsets of input data [32,34,39,46].…”

“…In the area of drug discovery, active learning has been shown to efficiently derive high-performance prediction models based on small subsets of input data [32,34,39,46]. Furthermore, actively trained models not only reached significantly higher hit rates compared to experimental standards which frequently remain below 1 % in cases of unbiased chemical libraries [34,39,40,[47][48][49], but such models also contributed to successful identification of novel bioactive compounds [33,36,42] and cancer rescue mutants of p53 [31].…”

Machine Learning in Chemical Industry

“…Machine learning techniques such as ANN can be used to automate this process by learning classification models from training compounds of each class [9]. It is important that the input contains all the latent feature information of the chemical compound.…”