Natural hydrogen from deep underground could be an important building block for the sustainable energy system of the future, but it is currently still difficult to predict where and at what depth elevated concentrations are located. New study results from the Department of Geology at the University of Vienna could make such predictions easier in the future. So-called "fairy circles" – round patches with vegetation damage – could be helpful indicators. This is because these "fairy circles" subside due to the seepage of natural hydrogen. According to the study, the deeper the hydrogen source underground and the higher its pressure, the larger the "fairy circle" on the Earth's surface. The study was recently published in the journal Geology.

A research team has developed a novel weakly supervised deep learning method that reconstructs spectral data from inexpensive RGB images, eliminating the need for manual labeling.

Recent advances in spatial and single-cell omics have significantly revolutionized biomarker discovery in tumor immunotherapy by addressing critical challenges such as tumor heterogeneity, immune evasion, and variability within the tumor microenvironment (TME). Immunotherapeutic strategies, including immune checkpoint inhibitors and adoptive T-cell transfer, have demonstrated promising clinical outcomes; however, their efficacy is limited by low response rates and the incidence of immune-related adverse events (irAEs). Therefore, the identification of reliable biomarkers is essential for predicting treatment efficacy, minimizing irAEs, and facilitating patient stratification. Spatial omics integrates molecular profiling with spatial localization, thereby providing comprehensive insights into the cellular organization and functional states within the TME. By elucidating the spatial patterns of immune cell infiltration and tumor heterogeneity, this approach enhances the prediction of therapeutic responses. Similarly, single-cell omics enables high-resolution analysis of cellular heterogeneity by capturing transcriptomic, epigenomic, and metabolic signatures at the single-cell level. The integrated application of spatial and single-cell omics has enabled the identification of previously undetected biomarkers, including rare immune cell subsets implicated in resistance mechanisms. In addition to spatial transcriptomics (ST), this technological landscape also includes spatial proteomics (SP) and spatial metabolomics, which further facilitate the study of dynamic tumor-immune interactions. Multi-omics integration provides a comprehensive overview of biomarker landscapes, while the rapid evolution of artificial intelligence (AI)-based approaches enhances the analysis of complex, multidimensional datasets to ultimately enhance predictive potential and clinical utility. Despite substantial progress, several challenges remain in the context of standardization, data integration, and real-time monitoring. Nevertheless, the incorporation of spatial and single-cell omics into biomarker research holds transformative potential for advancing personalized cancer immunotherapy. These emerging strategies pave the way for the development of innovative diagnostic and therapeutic interventions, thereby enabling precision oncology and improving treatment outcomes across a wide range of tumor profiles.

This review aims to provide a comprehensive overview of the integration of spatial omics with single-cell omics in the discovery of biomarkers for tumor immunotherapy. Specifically, it examines the strategies by which these emerging technologies address the challenges related to tumor heterogeneity, immune evasion, and the dynamic nature of the TME. By elaborating on the principles, applications, and clinical potential of these technologies, this review also critically evaluates their limitations, challenges, and the current gaps in clinical translation.

Identifying embryos with the highest likelihood of successful implantation is a critical component of the in vitro fertilization (IVF) process. Visual assessments are limited by the subjectivity of embryologists, making consistent evaluation of embryo health challenging with traditional methods. Recent advances in artificial intelligence (AI)—particularly in computer vision and deep learning—have enabled the automated analysis of embryo morphology images, reducing subjectivity and improving evaluation efficiency. Through an extensive literature search using keywords such as “embryo health assessment” and “artificial intelligence,” the present review focuses on AI-driven approaches for automated embryo evaluation. It examines AI techniques applied to embryo assessment across the early development, blastocyst, and full developmental stages. This review indicated the promising potential of AI technologies in enhancing the precision, consistency, and speed of embryo selection. AI models have been reported to outperform manual evaluations across several parameters, offering promising opportunities to improve success rates and operational efficiency in reproductive medicine. Additionally, this review discusses the current limitations of AI implementation in clinical settings and explores future research directions. Overall, the review provides insight into AI’s growing role in advancing embryo selection and highlights the path toward fully automated evaluation systems in assisted reproductive technology.

A research team demonstrates that combining unmanned aerial vehicle (UAV) multispectral imagery with a three-dimensional radiative transfer model (3D RTM) and machine learning can overcome shadow interference, enabling precise, orchard-scale mapping of leaf and canopy chlorophyll content.

A research team has demonstrated that consumer-grade digital cameras, paired with machine learning, can rapidly and accurately identify waterlogging-tolerant chrysanthemum varieties.