New exploratory research reveals dramatic mesozooplankton differences across bays and within single embayments less than 1 km apart, challenging uniform plankton monitoring practices.

Millions of people worldwide live with periodontitis, a chronic gum disease that can progressively destroy the tissues and bone supporting the teeth despite conventional treatment. Seeking a more comprehensive solution, researchers developed a novel DNA-based nanoplatform designed to deliver multiple therapeutic agents in a single package. By targeting the complex biological processes that drive the disease, this innovative approach represents a promising step toward next-generation therapies for chronic oral inflammation.

Researchers from Chinese People’s Liberation Army General Hospital, China, explored the use of techniques such as ‘intraoperative MRI (iMRI)’, ‘neuronavigation’, and ‘neuromonitoring’ to remove a rare and difficult-to-treat brain tumor subtype called ‘IDH-mutant low-grade corpus callosum glioma’. These techniques enabled safer and more complete tumor removal by helping surgeons delineate the tumor in real time and avoid damage to surrounding tissue, thereby improving survival by more than three years compared to conventional surgery.

A new study demonstrates that ramped pyrolysis/oxidation (RPO) can directly distinguish the sources of organic carbon transported by rivers after earthquakes and extreme rainstorms. By tracing carbon from vegetation, soils, and sedimentary rocks, the technique reveals how different natural hazards reshape carbon cycling and provides a powerful new tool for investigating climate-carbon feedbacks in rapidly changing mountain landscapes. The study was recently published in Science China Earth Sciences.

How can the long-term restoration effects of China’s Shan-Shui Initiative, recognized by the United Nations as one of the first World Restoration Flagships, be scientifically assessed? A research team from Chang’an University used long-term remote sensing and climate datasets from 2011 to 2024 to conduct a quantitative assessment of 25 early Shan-Shui Initiative across China. By developing an improved Remote Sensing Ecological Index, or kRSEI, the study found that vegetation greening does not always translate into broader improvements in satellite-observed ecological conditions. It also identified region-specific restoration pathways for forests, grasslands and croplands, and highlighted soil moisture as a key factor associated with the long-term stability of restoration outcomes. The findings provide a remote-sensing-based assessment framework that may support the Beautiful China initiative and offer a quantitative reference for evaluating large-scale ecological restoration projects.

Researchers have developed a controllable hydrothermal synthesis strategy for spherical cerium oxide (CeO₂) abrasives with tunable particle sizes from 32 to 531 nm while maintaining excellent sphericity and dispersity. Mechanistic studies reveal that CeO₂ formation occurs via nitrateinduced oxidation of Ce³⁺, with subsequent precipitation coupled to acetate decomposition thermodynamically driving the reaction forward. Chemical mechanical polishing (CMP) tests show that 343 nm CeO₂ achieves the highest material removal rate (73.7 nm·min⁻¹) with low surface roughness (Ra = 0.31 nm). The hydrothermally synthesized CeO₂ outperforms commercial abrasives in both short and longterm polishing owing to its uniform morphology.

In data center networks, the growth rate of switch buffer size is much smaller than that of network bandwidth capacity. This results in frequent network congestion, which requires congestion control. Currently, the classification of end-to-end congestion control protocols is based on single dimensions such as time or congestion arbitration subjects.

A Chinese research team has achieved a breakthrough in improving the training efficiency of Graph Neural Networks (GNNs). They introduced an innovative architecture named "Decentralized Hypercube Collaborative Framework," addressing long-standing challenges such as high memory overhead, low computational efficiency, and underutilized hardware resources in traditional Graph Convolutional Network (GCN) training.