Shallow Representation Learning via Kernel PCA Improves QSAR Modelability

Abstract

Linear models offer a robust, flexible, and computationally efficient set of tools for modeling quantitative structure activity relationships (QSAR), but have been eclipsed in performance by non-linear methods. Support vector machines (SVMs) and neural networks are currently among the most popular and accurate QSAR methods because they learn new representations of the data that greatly improve modelability. In this work we use shallow representation learning to improve the accuracy of L1 regularized logistic regression (LASSO) and meet the performance of Tanimoto SVM. We embedded chemical fingerprints in Euclidean space using Tanimoto (aka Jaccard) similarity kernel principal components analysis (KPCA), and compared the effects on LASSO and SVM model performance for predicting the binding activities of chemical compounds against 102 virtual screening targets. We observed similar performance and patterns of improvement for LASSO and SVM. We also empirically measured model training and cross validation times to show that KPCA used in concert with LASSO classification is significantly faster than linear SVM over a wide range of training set sizes. Our work shows that powerful linear QSAR methods can match nonlinear methods, and demonstrates a modular approach to non-linear classification that greatly enhances QSAR model prototyping facility, flexibility, and transferability.

Machine learning in chemoinformatics and drug discovery

“…used PCA to decorrelate features for prediction of estrogen receptor binding [69]. More recently, Rensi and Altman demonstrated performance improvements over LASSO regression for predicting activity against a broad set of pharmacological protein targets using kernel principal components analysis, and nonlinear variant of PCA [70]. Another popular regression method is partial least squares (PLS), which couples dimensionality reduction with multivariate regression to transform predictors into uncorrelated variables that are maximally correlated with the activity or property of interest.…”

Implicit-descriptor ligand-based virtual screening by means of collaborative filtering

“…Some works, rather than convert to problem to binary, opt to model the binding affinity value as a multivariate regression problem [46]. Even so, these works often compare the regression output to a static threshold to evaluate the accuracy of the methods resulting in a binary decision problem.…”

Drug Discovery of Spinal Muscular Atrophy (SMA) from the Computational Perspective: A Comprehensive Review

“…To date, AI has been extensively adopted to support healthcare services and research. Virtual screening [248], quantitative structure-activity relationship (QSAR) [249], de novo drug design [250,251], drug repurposing [252] and chemical space visualization [253] utilized ML extensively to reduce the gap in the conventional methods in drug discovery, while DL shows promise in proposing potent drug candidates using their properties and toxicity risks [9]. Uptake from the pharmaceutical industry is still lagged, especially for rare diseases.…”