New research brings machine‑learning‑based physics a step closer to solving real engineering challenges
Peer-Reviewed Publication
Updates every hour. Last Updated: 21-Apr-2026 06:16 ET (21-Apr-2026 10:16 GMT/UTC)
A mathematics professor at The University of Manchester has developed a novel machine-learning method to detect sudden changes in fluid behaviour, improving speed and cost of identifying these instabilities and overcoming one of the major obstacles faced when using machine learning to simulate physical systems.
A review in Glycoscience & Therapy summarizes advances in selective editing of N-glycan signals on living cell surface via the LCS‑NGS technology platform. It focuses on enzyme‑based strategies (glycosyltransferases and glycosidases) that enable precise, real‑time N‑glycan manipulation while preserving cell viability. Applications in tumor targeting, immune regulation, and cell therapy are highlighted. Importantly, glycan signals as a"biological quick recognition (BioQR)" hypothesis proposed—that glycan signals act as rapid markers for species, tissues, and disease states. Current limitations and future directions are also discussed.
Researchers from China developed the first high-density 10K SNP array for wax gourd using genotyping by target sequencing (GBTS) technology, comprising 10,722 genome-wide SNPs distributed across the genome, including 278 associated with functional trait loci.
First development of a RT-ERA/CRISPR-Cas12a platform for rapid and visual detection of Epizootic Hemorrhagic Disease Virus (EHDV), enabling on-site diagnosis within 1 h without specialized equipment.
- Prof. Sangwon Seo’s team successfully achieved enantioselective synthesis, selectively producing only the desired mirror-image molecule using nickel—a common metal—instead of expensive noble metals - Expected to make a breakthrough contribution to drug development that fundamentally blocks side effects and to the high-value precision chemical industry - Published in the top-tier international chemistry journal Angewandte Chemie – International Edition
Dragonflies have evolved special light-sensing proteins that let them see deeper red light than most animals. Researchers have now discovered that the mechanism of red vision is shared with humans and this ability comes from small molecular changes that could inspire new biomedical technologies.
Nagoya University researchers demonstrated that native soil bacteria, when treated with decoy molecules, can degrade non-native compounds, including persistent pollutants such as dioxins, without genetic modification.
NTU Singapore scientists have identified a fat-producing enzyme (GPAT) in brain cells that amplifies the damage caused by α-synuclein, the protein linked to Parkinson's disease. GPAT delivers a "double hit" — impairing cells' energy-producing machinery while increasing the protein's toxicity. Reducing GPAT activity led to less brain cell damage in lab models. The findings point to a potential new treatment target for a disease that currently has no cure.