Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
and JavaScript.
Autoencoders are versatile tools for molecular informatics with the opportunity for advancing molecule and drug design. In this Review, the authors highlight the active areas of development in the field and explore the challenges that need to be addressed moving forward.
Using deep learning methods to study gene regulation has become popular, but designing accessible and customizable software for this purpose remains a challenge. This work introduces a computational toolkit called EUGENe that facilitates the development of end-to-end deep learning workflows in regulatory genomics.
The accurate prediction of molecular spectra is essential for substance discovery and structure identification, but conventional quantum chemistry methods are computationally expensive. Now, DetaNet achieves the accuracy of quantum chemistry while improving the efficiency of prediction of organic molecular spectra.
The capability of predicting stable materials is important to further accelerate the discovery of novel materials. In this Review, the authors discuss recent developments in machine learning techniques for assessing the stability of materials and highlight the opportunities in further advancing the field.
A pairwise binding comparison network (PBCNet) has been established for predicting the relative binding affinity among congeneric ligands, using a physics-informed graph attention mechanism with a pair of protein pocket-ligand complexes as input. PBCNet shows practical value in guiding structure-based drug lead optimization with speed, precision, and ease-of-use.
Inspired by the classic lock-and-key model and advances in equivariant deep network design, we present a structure-based drug design model, SurfGen, which uses two types of equivariant graph neural networks to learn on protein surfaces and geometric structures to directly design small-molecule drugs.
We introduce STAligner — a graph neural network-based tool for the integration of multiple spatial transcriptomics datasets by generating batch effect-corrected embeddings, thereby enabling consensus spatial domain identification and accurate 3D tissue reconstruction.
Programmability is crucial in noisy intermediate-scale quantum computing, facilitating various functionalities for practical applications. An arbitrary programmable time-bin-encoded quantum boson sampling device has been developed, specifically tailored for potential drug discovery.
A guided diffusion model pushes the boundaries of de novo molecular design, extensively exploring the chemical space and generating chemical compounds that satisfy custom target criteria.
A recent study proposes a unified framework that can compare different measures for quantifying the statistics of pairwise interactions in data from complex dynamical systems.
Deep learning approaches have potential to substantially reduce the astronomical costs and long timescales involved in drug discovery. KarmaDock proposes a deep learning workflow for ligand docking that shows improved performance against both benchmark cases and in a real-world virtual screening experiment.
Different cells can have very different three-dimensional morphologies. We present the computational framework u-signal3D that calculates the spatial scales at which molecules are organized on the surfaces of heterogeneously shaped cells, enabling high-throughput analyses and subsequent machine learning applications.
A reinforcement-learning-based framework is proposed for assisting urban planners in the complex task of optimizing the spatial design of urban communities.
Graph neural networks (GNNs) present a promising route for machine learning of solid-state materials’ properties, but methods capable of directly predicting defect properties from ideal, defect-free structures are needed. A GNN developed for direct defect property predictions enables high-throughput screening of redox-active oxides for energy applications and beyond.
The dissipation and bending of light waves by atmospheric turbulence adversely affects infrared imaging, leading to grayscale drift, distortion, and blurring. A deep learning method has been developed to both extract the two-dimensional atmospheric turbulence strength fields and obtain clear and stable images from turbulence-distorted infrared images.
A hierarchical Bayesian method identifies cell-type specific changes in gene regulatory circuits in disease by integrating single-cell and three-dimensional measurements of the genome.
While the adherence to fairness constraints has become common practice in the design of algorithms across many contexts, a more holistic approach should be taken to avoid inflicting additional burdens on individuals in all groups, including those in marginalized communities.