Network image

Graph neural networks

Graph neural networks are a class of deep learning methods that can be used to model a variety of systems.

Announcements

  • image of two scientists working

    Topics soon to be featured include: Single-molecule fluorescence resonance energy transfer, Near-infrared-II fluorescence imaging, Quantitative text analysis and Freeze casting.

  • NRMP logo

    Never miss a Primer! The editors will be posting our newest content along with information about conferences and new developments in methods research.

Advertisement

  • Biocarbons are carbonaceous solids derived from renewable and sustainable feedstocks through thermochemical conversion at high temperatures. In this Primer, Mohanty et al. discuss feedstock selection, pyrolysis techniques and post-modification strategies, as well as waste reduction and the economic impact of biocarbons.

    • Amar K. Mohanty
    • Singaravelu Vivekanandhan
    • Manjusri Misra
    Primer
  • Graph neural networks are a class of deep learning methods that can model physical systems, generate new molecules and identify drug candidates. This Primer introduces graph neural networks and explores how they are applied across the life and physical sciences.

    • Gabriele Corso
    • Hannes Stark
    • Regina Barzilay
    Primer
  • Structures of surfaces and thin films can be investigated by performing X-ray diffraction under grazing incidence conditions. This Primer explores how grazing incidence X-ray diffraction is used to obtain crystallographic information, including in situ characterization, data collection, analysis and visualization, across a range of applications.

    • Oliver Werzer
    • Stefan Kowarik
    • Roland Resel
    Primer
  • Artificial molecular pumps are synthetic machines capable of performing complex tasks on a molecular scale. In this Primer, Zhang et al. discuss the design features and underlying fundamental physical principles of artificial molecular pumps.

    • Long Zhang
    • Huang Wu
    • J. Fraser Stoddart
    Primer
  • Fresh water can be produced from saline water using pressure-driven membrane desalination. This Primer explores how reverse osmosis and nanofiltration are used as energy-efficient desalination methods, with a focus on membrane development, characterization and performance modelling.

    • Weifan Liu
    • Joshua L. Livingston
    • Shihong Lin
    Primer
  • Brillouin microscopy is a non-contact method used for mechanical probing of cells and tissues. In this Primer, Kabakova et al. provide a comprehensive overview of the methods and applications of Brillouin microscopy.

    • Irina Kabakova
    • Jitao Zhang
    • Giuliano Scarcelli
    Primer
  • Bayesian optimization is a promising approach towards a more environmentally friendly chemical synthesis, in line with the Sustainable Development Goals. It can aid chemists to explore vast chemical spaces and find green reaction conditions with few experiments, decreasing resource consumption and waste generation while reducing discovery timelines and costs.

    • Elena Braconi
    Comment
  • To improve early-stage research in the field of RNA lipid nanoparticles, there are several best practices to be considered for the collection, interpretation and reporting of characterization data.

    • Omar F. Khan
    Comment
  • To ensure a sustainable future and combat food scarcity, we must boost agricultural productivity, improve climate resilience and optimize resource usage. There is untapped potential for dense wireless sensor networks in agriculture that can increase yields and support resilient production when linked to smart decision and control systems.

    • Peter G. Steeneken
    • Elias Kaiser
    • Marie-Claire ten Veldhuis
    Comment
  • New nanomaterials are being developed for efficient biomolecule delivery to plants. However, detection and quantification of plant cell entry are challenging and currently rely on subjective methods that lack proper controls. The necessary considerations of performing nanoparticle-mediated delivery in plants and how to accurately quantify delivery efficiency are discussed.

    • Gozde S. Demirer
    Comment
  • Laboratory hardware is often custom made or significantly modified. To improve reproducibility, it is imperative that these novel instruments are properly documented. Increasing adoption of open source hardware practices can potentially improve this situation. This article explores how open licences and open development methodologies enable custom instrumentation to be reproduced, scrutinized and properly recorded.

    • Richard W. Bowman
    Comment