| Title: | Plasma proteomic signatures associated with ulcerative colitis risk in overweight/obesity individuals: a UK Biobank study |
| Journal: | Scientific Reports |
| Published: | 27 Aug 2026 |
| DOI: | https://doi.org/10.1038/s41598-026-67858-3 |
| URL: | http://dx.doi.org/10.1038/s41598-026-67858-3 |
Publication 20871
WARNING: the interactive features of this website use CSS3, which your browser does not support. To use the full features of this website, please update your browser.
Abstract
The association between overweight/obesity and ulcerative colitis (UC) has been observed in various cohorts, however, the underlying systemic proteomic profiles and their predictive potential within this specific population remain to be fully elucidated. This study aimed to identify plasma proteomic signatures associated with the obesity and UC to evaluate an exploratory risk assessment model using machine learning. Data from 52,445 UK Biobank participants with Olink proteomics data were analyzed. Stepwise multivariable logistic regression models were employed to assess associations, adjusting for age, sex, ethnicity, smoking status, alcohol consumption, medication use, baseline C-reactive protein (CRP) levels, and sleep parameters. Core candidate differentially expressed proteins (DEPs) were identified through a sequential filtering pipeline intersecting cross-sectional and prospective studies. A risk assessment model was developed using 10-fold cross-validation on a 1:1 propensity-score matched prospective study (n = 350). In this study, approximately 70% of participants in both the UC and pre-UC cohorts were overweight or obesity and exhibited elevated CRP levels. Seven core candidate biomarkers (IL2RA, TNF, TGFA, CCL20, CKB, CXCL9, and IGFBP2) were significantly dysregulated in both overweight/obesity and UC. In a refined validation cohort (n = 37,478), all seven proteins maintained highly significant independent associations with UC risk after full adjustment for all clinical and lifestyle confounders (all p < 0.01). Functional enrichment analysis highlighted pathways involving cytokine-cytokine receptor interaction and PI3K-Akt signaling. Among the machine learning algorithms, the logistic regression model demonstrated the most stable performance, achieving a moderate but consistent discriminative ability with an AUC of 0.715 in the training set and 0.714 in the independent testing set. This study identifies a robust plasma proteomic signature independently associated with UC risk in individuals with overweight and obesity. These findings provide an exploratory molecular framework for risk stratification. While these proteins serve as promising candidate biomarkers, their moderate predictive performance suggests they reflect a persistent systemic pro-inflammatory state that precedes clinical UC onset. Further studies are warranted to validate these signatures in external cohorts.</p>
4 Authors
- Shujian Wu
- Yifen Lin
- Hualin Duan
- Aimin Xu
1 Application
| Application ID | Title |
|---|---|
| 217792 | Multi-omics analyses for diabetes and its complications |
Enabling scientific discoveries that improve human health