Predictive markers articles from across Nature Portfolio

To capture the changing rather than one-off risk of type 1 diabetes (T1D), we developed an artificial intelligence-driven microRNA-based multi-context (5,983 samples from 7 countries) dynamic risk score (DRS). We validated that this DRS enabled T1D discrimination, misdiagnosis correction and therapeutic response prediction, establishing a personalized, adaptive framework for T1D risk monitoring and intervention.

This study shows that structural complexity of cortical and subcortical areas, measured from standard anatomical MRI, can predict Parkinson’s patient response to deep brain stimulation using advanced hypergraph modeling techniques.

Differences in the Pace of Aging are important for many health outcomes but difficult to measure. Here the authors describe the Dunedin Pace of Aging Calculated from NeuroImaging measure, an approach that uses a single brain image to measure how fast a person is aging and can help predict mortality or the risk of developing chronic disease.

Authors report MagNet, a plasma extracellular vesicle (EV) enrichment strategy using magnetic beads. Proteomic interrogation of this plasma EV fraction enables the detection of proteins that are beyond the dynamic range of mass spectrometry of unfractionated plasma.

Analyzing readouts of inflammaging across four cohorts, Franck and colleagues identify strong variation and observe that inflammaging, in its known form, primarily emerges in industrialized—but not nonindustrialized—populations.

To capture the changing rather than one-off risk of type 1 diabetes (T1D), we developed an artificial intelligence-driven microRNA-based multi-context (5,983 samples from 7 countries) dynamic risk score (DRS). We validated that this DRS enabled T1D discrimination, misdiagnosis correction and therapeutic response prediction, establishing a personalized, adaptive framework for T1D risk monitoring and intervention.

Tissue-agnostic cancer therapies promise to revolutionize oncology by targeting molecular drivers. Sledge et al.’s study of nearly 300,000 tumors found 21.5% with tissue-agnostic indications. Despite nine FDA approvals, real-world implementation challenges persist. Progress depends on universal genomic testing, an oncogenomic-savvy workforce, innovative trials, updated regulations, and real-world evidence to maximize potential.