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.
Machine learning is swiftly infiltrating many areas within the healthcare industry, from diagnosis and prognosis to drug development and epidemiology, with significant potential to transform the medical landscape.
Rapid progress in machine learning is enabling opportunities for improved clinical decision support. Importantly, however, developing, validating and implementing machine learning models for healthcare entail some particular considerations to increase the chances of eventually improving patient care.
Highly quantitative, robust, single-cell analyses can help to unravel disease heterogeneity and lead to clinical insights, particularly for complex and chronic diseases. Advances in computer vision and machine learning can empower label-free cell-based diagnostics to capture subtle disease states.
At the recent Artificial Intelligence Applications in Biopharma Summit in Boston, USA, a panel of scientists from industry who work at the interface of machine learning and pharma discussed the diverging opinions on the past, present and future role of AI for ADME/Tox in drug discovery and development.
This Comment describes some of the common pitfalls encountered in deriving and validating predictive statistical models from high-dimensional data. It offers a fresh perspective on some key statistical issues, providing some guidelines to avoid pitfalls, and to help unfamiliar readers better assess the reliability and significance of their results.
By measuring the photocurrent from illuminated Weyl semimetals, an optical signature of topological properties arising from Weyl fermions has been revealed, highlighting nonlinear optical effects and applications of Weyl semimetals.
Neural probes that mimic the subcellular structural features and mechanical properties of neurons assimilate across several structures of the brain to provide chronically stable neural recordings in a mouse model.
This Perspective describes the application of machine learning models in the design, synthesis and characterisation of molecules at different stages in the drug discovery and development process.
An investigation of the structural and transport properties of bilayer graphene as a function of the twist angle between the layers reveals atomic-scale reconstruction for twist angles smaller than a critical value.
Facile absorption and desorption of hydrogen at palladium surfaces provides a way to define how metal–solute interactions impact properties relevant to energy storage, catalysis and sensing. In situ X-ray diffraction has now been used to track both hydrogen absorption and desorption in palladium nanocrystals.
Charge-transfer states with comparable recombination and charge-splitting rates are shown to be a key ingredient for donor–acceptor organic blends that perform well in both light-emitting and photovoltaic applications.
Molten salts are used as a reaction medium to protect carbide, nitride and boride powders from oxidation during high-temperature synthesis in air, thus avoiding the need to carry out these processes in a vacuum or inert environment.
High mobility and high carrier density are found in the Weyl semimetal NbAs. This is attributed to the low dissipation of disorder-tolerant Fermi arcs.
Sensing hydrogen by the change in plasmonic response upon metal hydride formation is safe, but trace gas poisoning and low sensitivity can occur. Here, a PdAu alloy/polymer sensor is poison resistant and can sense 3 ppm H2 with a response time of 1 s.
Although anionic redox in Li- and Na-rich transition metal oxides can enhance energy density of rechargeable batteries, anionic capacity is partly irreversible in discharge. A unified picture to clarify this irreversibility and to improve cycling performance is proposed.
The structural foundation of self-assembled peptide materials is typically the β-sheet. Here the authors describe peptides made of three natural amino acids that self-assemble into helical-like superstructures with enhanced mechanical rigidity.
Neural probes mimicking the size and mechanical properties of neurons interpenetrate the brain tissue, allowing stable single-unit recordings from implantation up to at least three months, and acting as scaffolds for the migration of new-born neurons.