The abstract is read at the publisher; the summary is JClub's.
Network-guided plasma protein panels across 23 human diseases achieve mean AUROC of 0.741 (disease vs healthy) and 0.736 (disease vs diseases), outperforming a shared-protein baseline in 7/23 per task.
- Why it matters: Large plasma proteomics datasets need compact, interpretable protein panels that stay accurate in independent cohorts, but this cross-cohort generalization is difficult.
- What they did: The framework links the Human Disease Blood Atlas to the UK Biobank Pharma Proteomics Project using 1154 shared quantified proteins, then builds disease-specific panels for two classification tasks across 23 diseases.
- The result: Panels integrate panel membership with differential protein levels and disease-associated co-expression modules to surface disease-enriched candidates from disease-shared proteins linked to multiple conditions.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.