Researchers have developed an AI-assisted method for muon scattering tomography that replaces manual parameter tuning with reinforcement learning. An AI agent evaluates reconstruction quality and automatically adjusts key PoCA settings in a closed loop, yielding clearer and more structurally faithful images within challenging dense and shielded environments. Across four experiments, the method demonstrated substantial average gains over the standard PoCA approach, including a 307% improvement in shape overlap accuracy, a 79% increase in structural similarity, and an 8.4% rise in peak signal-to-noise ratio. It also showed overall improvement against an iterative MLSD baseline. The study appears in Nuclear Science and Techniques 37, 81 (2026). https://doi.org/10.1007/s41365-026-01894-6

The research team led by Hanmin Huang and Bangkui Yu at the University of Science and Technology of China developed a palladium-catalyzed diastereoselective and enantioselective cascade cyclization strategy, achieving the modular synthesis of chiral nitrogen-bridged ring skeletons. Using readily available salicylaldehyde and aminodiene as starting materials, and based on the team's previously developed strategy of "in-situ generation of three-membered ring palladium active intermediates from aldehydes and amines," the bridged oxazole bicyclic compounds were constructed with high diastereoselectivity through a continuous cyclization process. This method exhibits excellent substrate universality, providing an efficient and precise route for synthesizing drug molecules with complex three-dimensional structures. The article was published as an open access Communication in CCS Chemistry, the flagship journal of the Chinese Chemical Society.

Study involving researchers from a FAPESP-supported center presents a new molecular architecture based on zirconium metal-organic frameworks (MOFs) designed for efficiently degrading emerging water contaminants.

A beneficial bacterium developed at the University of Delaware cuts turfgrass dollar spot disease by 43.6% – but only with direct leaf application. Their new study details the need for viable, targeted biocontrol strategies in turf management.

Researchers from the University of Tartu Institute of Physics have developed a novel method for enhancing the quality of three-dimensional images by increasing the depth of focus in holograms fivefold after recording, using computational imaging techniques. The technology enables improved performance of 3D holographic microscopy under challenging imaging conditions and facilitates the study of complex biological structures. 

Weight loss is common in Parkinson’s disease, but its biological basis has been unclear. Researchers at Fujita Health University show that this loss reflects reduced body fat, not muscle, along with a shift in energy metabolism. Patients exhibited impaired carbohydrate metabolism, mitochondrial dysfunction, and increased fat breakdown with ketone body production. These changes were most pronounced in thinner patients and those with more advanced disease, revealing a hidden energy crisis in Parkinson’s disease.

Understanding the thermodynamic basis of ligand recognition by G-protein-coupled receptors (GPCRs), especially in terms of enthalpy-entropy compensation, is crucial for drug design and development. However, such thermodynamic parameters for GPCRs have been largely unexplored. To address this gap, researchers investigated the binding enthalpy and entropy characteristics of the histamine H₁ receptor with doxepin and its individual isomers, revealing new insights for improving drug selectivity and reducing side effects.

Controlled manipulation of the fibers that are as thin as or even thinner than human hair is a real challenge. Despite the technology development, the precise and reversible change of the microfibers' orientation, like the tweezers, is not easy. The interdisciplinary team of researchers from the Institute of Physical Chemistry, Polish Academy of Sciences, has recently developed a way to control the shape of microfibers with electricity. This brings us closer to a novel technical solution in micromechanics and soft robotics. Their recent work, published in the Nature Communications journal, demonstrates the first proof-of-concept results on the motion of pristine carbon fibers caused by asymmetric electrochemical processes occurring in the material.

On 2 July, 2025, the China-led Einstein Probe (EP) space telescope detected an exceptionally bright X-ray source whose brightness varied rapidly during a routine sky survey. Its unusual signal immediately set it apart from ordinary cosmic sources, triggering rapid follow-up observations by telescopes worldwide.

This research was coordinated by the EP Science Center of the National Astronomical Observatories, Chinese Academy of Sciences (NAOC), with participation from multiple research institutions in China and abroad. Astrophysicists from the Department of Physics at The University of Hong Kong (HKU), who are integral members of the EP scientific team, worked together with the broader collaboration to interpret the event, proposing that it may mark the moment when an intermediate-mass black hole tears apart and consumes a white dwarf star. If confirmed, this would be the first observational evidence of such an extreme black hole “feeding” process. The findings have been published as a cover article in Science Bulletin.