Search

Generative molecular design (GMD) is an increasingly popular strategy for drug discovery, using machine learning models to propose, evaluate and optimize chemical structures against a set of target design criteria. We present the ATOM-GMD platform, a scalable multiprocessing framework to optimize many parameters simultaneously over large populations of proposed molecules. ATOM-GMD uses a junction tree variational autoencoder mapping structures to latent vectors, along with a genetic algorithm operating on latent vector elements, to search a diverse molecular space for compounds that meet the design criteria. We used the ATOM-GMD framework in a lead optimization case study to develop potent and selective histamine H1 receptor antagonists. We synthesized 103 of the top scoring compounds and measured their properties experimentally. Six of the tested compounds bind H1 withKi’s between 10 and 100 nM and are at least 100-fold selective relative to muscarinic M2 receptors, validating the effectiveness of our GMD approach.

Drug-induced liver injury (DILI) is the most common cause of acute liver failure and a frequent reason for withdrawal of candidate drugs during preclinical and clinical testing. An important type of DILI is cholestatic liver injury, caused by buildup of bile salts within hepatocytes; it is frequently associated with inhibition of bile salt transporters, such as the bile salt export pump (BSEP). Reliable in silico models to predict BSEP inhibition directly from chemical structures would significantly reduce costs during drug discovery and could help avoid injury to patients. Unfortunately, models published to date have been insufficiently accurate to encourage wide adoption. We report our development of classification and regression models for BSEP inhibition with substantially improved performance over previously published models. Our model development leveraged the ATOM Modeling PipeLine (AMPL) developed by the ATOM Consortium, which enabled us to train and evaluate thousands of candidate models. In the course of model development, we assessed a variety of schemes for chemical featurization, dataset partitioning and class labeling, and identified those producing models that generalized best to novel chemical entities. Our best performing classification model was a neural network with ROC AUC = 0.88 on our internal test dataset and 0.89 on an independent external compound set. Our best regression model, the first ever reported for predicting BSEP IC50s, yielded a test set $R^2 = 0.56$ and mean absolute error 0.37, corresponding to a mean 2.3-fold error in predicted IC50s, comparable to experimental variation. These models will thus be useful as inputs to mechanistic predictions of DILI and as part of computational pipelines for drug discovery.

Hidden within our rapidly growing global streams of business, scientific, and communications data, is information that can reduce the global danger of nuclear weapons proliferation and use. Significant advances in machine learning, the explosion of available data and the ability of high performance computing that can be used to process massive amounts of data has the potential to provide a leap-ahead in the ability to detect proliferation and verify treaty commitments.

Machine learning is finding increasingly broad applications in the physical sciences. This most often involves building a model relationship between a dependent, measurable output, and an associated set of controllable, but complicated, independent inputs. We present a tutorial on current techniques in machine learning—a jumping-off point for interested researchers to advance their work. We focus on deep neural networks with an emphasis on demystifying deep learning. We begin with background ideas in machine learning and some example applications from current research in plasma physics. We discuss supervised learning techniques for modeling complicated functions, beginning with familiar regression schemes, and then advancing to more sophisticated deep learning methods. We also address unsupervised learning and techniques for reducing the dimensionality of input spaces. Along the way, we describe methods for practitioners to help ensure that their models generalize from their training data to as-yet-unseen test data. We describe classes of tasks—predicting scalars, handling images, and fitting time-series—and prepare the reader to choose an appropriate technique. We finally point out some limitations to modern machine learning and speculate on some ways that practitioners from the physical sciences may be particularly suited to help.