Researchers from Aarhus University have developed a new way of thinking about biodiversity. Instead of measuring only the species we can observe today, their new concept aims to estimate how much biodiversity a landscape is actually capable of supporting. The approach could help explain why some conservation measures succeed while others deliver disappointing results.

A review in iMetaMed examines how gut microbial metabolites can promote or restrain colorectal cancer and influence antitumor immunity. It brings together evidence on short-chain fatty acids, bile acids and other microbial products, alongside approaches involving probiotics, prebiotics and engineered bacteria. The authors also assess microbial biomarkers for non-invasive detection, emphasizing context-dependent effects, unresolved causality, safety concerns and the need for standardized validation before these approaches can guide clinical care.

A study in iMetaMed presents DECEPTICONx, a computational framework for estimating detailed immune and stromal cell compositions from tumor RNA sequencing data. The researchers combined reference profiles from 535,665 single cells with more than 11,000 bulk samples across 30 cancer types, mapping 102 cell subtypes. The atlas identifies associations with survival and immunotherapy response, while its lack of spatial information and the need for mechanistic validation remain important limitations.

In a multicenter cohort study involving 11 hospitals in China, researchers developed a machine learning model to estimate operative duration in posterior spinal deformity correction surgery before the procedure begins. By analyzing clinical and laboratory data from 162 patients, the team evaluated ten algorithms and identified LightGBM as the most consistent performer. Seven preoperative variables—including age, ASA class, and several laboratory markers—were found to contribute most strongly to the prediction. Interpretability analyses using SHAP helped clarify how each factor influenced model output. The resulting tool, implemented as a web-based calculator, may support individualized surgical planning and more efficient allocation of operating-room resources.

Bizard is an open-source resource that brings biomedical visualization scripts, example datasets and tutorials together in one searchable platform. The iMetaMed study describes more than 750 rendered graphics across over 70 chart types, bilingual documentation and community contributions, while identifying limitations in language support and clinical application.