Key points from article :
Researchers have discovered that blood proteins can predict the risk of developing 67 diseases within the next 10 years, potentially offering a more accurate method than traditional clinical information. Conducted through the UK Biobank with 41,931 participants, the study measured around 3,000 plasma proteins, linking them to disease outcomes in the largest proteomics study to date.
This method covered a wide range of diseases including multiple myeloma, motor neuron disease, and pulmonary fibrosis. Sparse protein signatures, utilising the most important 5 to 20 proteins, outperformed models based solely on clinical information for these diseases.
For 52 of the diseases studied, protein-based models provided better predictions than clinical models that used blood assays. Even in the presence of polygenic risk scores for 23 diseases, the protein-based predictions were more accurate for all but one disease, breast cancer. The study used single-cell RNA sequencing from bone marrow of multiple myeloma patients, finding that four out of five predictor proteins were specifically expressed in plasma cells, consistent with their strong predictive power.
The research also included external validation using data from the EPIC-Norfolk study, confirming that proteins were superior to clinical models in six of the 67 diseases, supporting the generalisability of the findings.
Published in Nature Medicine, this study was led by researchers from Queen Mary University of London.
