Scientists have developed a hierarchical Transformer-based artificial intelligence model that predicts RNA splice sites across DNA sequences up to 100,000 base pairs long while preserving single-nucleotide resolution. The system captures long-range regulatory interactions that conventional approaches often miss and visualizes the sequence regions influencing its predictions. The approach achieved state-of-the-art performance on multiple benchmark datasets and may support future studies in genetic diseases, cancer, precision medicine, genomic medicine, and therapeutic development.

The Hebrew University of Jerusalem proudly congratulates two of its outstanding researchers on receiving prestigious European Research Council (ERC) Proof of Concept Grants for 2026. The grants support ERC-funded researchers in translating groundbreaking discoveries into innovative applications, technologies, and services with real-world impact.

A team of researchers at The University of Texas at El Paso has developed a way to 3D-print an essential battery component in nearly any shape. Their innovation could free engineers from the constraints of standard rechargeable battery sizes.

As one of the world’s most revered gourmet foods, some truffles can sell for more than €1,000 per kilogram, making cultivation a lucrative business.

A formal model recently published in the open-access Agricultural and Environmental Modelling journal describes the development of APODEMUS, which aims to improve risk assessments of pesticides for the wood mouse (Apodemus sylvaticus) in European agricultural landscapes. This computer simulation provides a foundational blueprint and a tool aimed at helping regulators accurately test the environmental safety of agricultural pesticides by simulating how they affect wild animal populations.

Researchers at Linköping University, Sweden, have now demonstrated how organic solar cells can become more efficient than previously thought possible. The key is to extend the time that electrons in the material remain excited, which leads to improved performance. The findings are published in the journal Nature Photonics.

Researchers developed a simple and cost-effective way to protect environmental monitoring equipment from wildfire damage. The redesigned data cabinets will keep electronics below their failure temperature during wildfires, helping maintain critical environmental data collection.

This study presents a hybrid dynamic best model selection algorithm for real-time building fire prediction using IoT-enabled smoke, temperature, and humidity sensors. Validated on a custom laboratory dataset and live sensor streams, the system dynamically selected the best machine-learning classifier and achieved 99.57% accuracy, 99.37% precision, 99.79% recall, and low error, improving adaptability and early fire detection.