Search

Abstract Traditional methods for identifying “hit” molecules from a large collection of potential drug-like candidates rely on biophysical theory to compute approximations to the Gibbs free energy of the binding interaction between the drug and its protein target. These approaches have a significant limitation in that they require exceptional computing capabilities for even relatively small collections of molecules. Increasingly large and complex state-of-the-art deep learning approaches have gained popularity with the promise to improve the productivity of drug design, notorious for its numerous failures. However, as deep learning models increase in their size and complexity, their acceleration at the hardware level becomes more challenging. Hyperdimensional Computing (HDC) has recently gained attention in the computer hardware community due to its algorithmic simplicity relative to deep learning approaches. The HDC learning paradigm, which represents data with high-dimension binary vectors, allows the use of low-precision binary vector arithmetic to create models of the data that can be learned without the need for the gradient-based optimization required in many conventional machine learning and deep learning methods. This algorithmic simplicity allows for acceleration in hardware that has been previously demonstrated in a range of application areas (computer vision, bioinformatics, mass spectrometery, remote sensing, edge devices, etc.). To the best of our knowledge, our work is the first to consider HDC for the task of fast and efficient screening of modern drug-like compound libraries. We also propose the first HDC graph-based encoding methods for molecular data, demonstrating consistent and substantial improvement over previous work. We compare our approaches to alternative approaches on the well-studied MoleculeNet dataset and the recently proposed LIT-PCBA dataset derived from high quality PubChem assays. We demonstrate our methods on multiple target hardware platforms, including Graphics Processing Units (GPUs) and Field Programmable Gate Arrays (FPGAs), showing at least an order of magnitude improvement in energy efficiency versus even our smallest neural network baseline model with a single hidden layer. Our work thus motivates further investigation into molecular representation learning to develop ultra-efficient pre-screening tools. We make our code publicly available at https://github.com/LLNL/hdbind .

Abstract Nerve agents have experienced a resurgence in recent times with their use against civilian targets during the attacks in Syria (2012), the poisoning of Sergei and Yulia Skripal in the United Kingdom (2018) and Alexei Navalny in Russia (2020), strongly renewing the importance of antidote development against these lethal substances. The current standard treatment against their effects relies on the use of small molecule-based oximes that can efficiently restore acetylcholinesterase (AChE) activity. Despite their efficacy in reactivating AChE, the action of drugs like 2-pralidoxime (2-PAM) is primarily limited to the peripheral nervous system (PNS) and, thus, provides no significant protection to the central nervous system (CNS). This lack of action in the CNS stems from their ionic nature that, on one end makes them very powerful reactivators and on the other renders them ineffective at crossing the Blood Brain Barrier (BBB) to reach the CNS. In this report, we describe the use of an iterative approach composed of parallel chemical and in silico syntheses, computational modeling, and a battery of detailed in vitro and in vivo assays that resulted in the identification of a promising, novel CNS-permeable oxime reactivator. Additional experiments to determine acute and chronic toxicity are ongoing.