Using Olink proteomics and machine learning, this study identified four immune-related plasma biomarkers (MGMT, SIT1, PRDX1, TRAF2) for Moyamoya disease diagnosis, with an XGBoost-based diagnostic model achieving the highest AUC value.
The Olink immunopanel identified 44 differentially expressed proteins in Moyamoya disease patients compared to healthy controls.
GO and KEGG enrichment analyses revealed significant enrichment of differentially expressed proteins in innate immune responses and NF-kB and MAPK signaling pathways.
Four proteins — MGMT, SIT1, PRDX1, and TRAF2 — were identified as potential biomarkers for Moyamoya disease.
The XGBoost machine learning model achieved the highest AUC value among the diagnostic models constructed using the four biomarkers.
TRAF2 and PRDX1 showed significant expression differences in Moyamoya disease patients in an independent external GEO dataset, providing validation of these biomarkers.
Potential therapeutic drugs for Moyamoya disease were predicted using pharmacogenomic databases and validated through molecular docking.
Moyamoya disease is a rare brain blood vessel disorder that currently requires an invasive procedure called digital subtraction angiography (DSA) for diagnosis. This research suggests that specific proteins found in blood plasma could serve as non-invasive diagnostic markers for this condition. The researchers used a technology called Olink proteomics to measure 92 immune-related proteins in blood samples from 88 Moyamoya disease patients and 88 healthy individuals, finding that 44 proteins were abnormally expressed in patients. This research suggests that four proteins in particular — MGMT, SIT1, PRDX1, and TRAF2 — are the most promising candidates as blood-based biomarkers for Moyamoya disease. Using artificial intelligence techniques including random forest and XGBoost machine learning models, the researchers built diagnostic tools based on these four proteins, with the XGBoost model performing best. Two of the biomarkers (TRAF2 and PRDX1) were further confirmed using an independent external database, adding confidence to the findings. The study also found that the abnormal proteins are heavily involved in immune and inflammatory signaling pathways, suggesting that immune dysfunction plays an important role in Moyamoya disease. If these findings are confirmed in larger studies, this research suggests it may eventually be possible to diagnose Moyamoya disease using a simple blood test rather than invasive procedures. Additionally, because the study identified specific proteins involved in the disease process and predicted potential drugs that might target them, these findings could also help guide the development of new treatments for this currently difficult-to-manage condition.
Zhang Z, Xu H, Hou Z, Ni H, Li Y, Zhou S, et al.. (2026). Identification of Immune Response-Related Proteomic Biomarkers in Moyamoya Disease Using Serum Olink Proteomics.. Journal of proteome research. https://doi.org/10.1021/acs.jproteome.6c00310