A processed dietary pattern is associated with gestational diabetes mellitus and altered APOM and SAA4 proteomic signatures in Miao women

Background This study integrated dietary pattern analysis with serum proteomics to identify diet-related molecular signatures associated with gestational diabetes mellitus (GDM) in Miao pregnant women. Methods A 1:1 matched case–control study was conducted among 78 Miao pregnant women (39 GDM cases and 39 controls). Dietary patterns were derived using principal component analysis. Serum proteomics was performed using data-independent acquisition mass spectrometry. Associations were evaluated using conditional logistic regression and linear mixed-effects models. LASSO penalized logistic regression was used for exploratory feature selection, followed by nested cross-validation to assess internal discriminative performance and an exploratory multivariable analysis. Results Higher adherence to a processed dietary pattern was associated with higher odds of GDM (OR = 2.33, 95% CI: 1.08–5.04). Proteomic analysis identified 34 candidate differentially expressed proteins enriched in lipid metabolism and inflammatory pathways. The processed dietary pattern was inversely associated with APOM (β = −0.308, p = 0.036) and showed a borderline inverse association with SAA4 ( β = −0.285, p = 0.052). A leave-one-matched-pair-out nested LASSO feature-selection analysis using all 344 quality-controlled proteins yielded an internal AUC of 0.953 (95% CI: 0.894–0.992). SAA4 was selected in all 39 outer iterations, whereas selection frequencies varied across other proteins. APOM was retained for subsequent exploratory integrative analysis based primarily on its association with the processed dietary pattern, PPI-hub status, and biological relevance. Conclusion A processed dietary pattern was associated with GDM and altered circulating APOM/SAA4 profiles in Miao pregnant women. Nested cross-validation provided an exploratory assessment of proteomic feature stability using all 344 quality-controlled proteins. The resulting internal discrimination should be interpreted cautiously because the upstream QC correction/normalization incorporated GDM group information and was completed before cross-validation. The APOM/SAA4–diet model was developed as a proof-of-concept analysis only and should not be interpreted as a clinically applicable prediction model. Larger independent and prospective cohorts are required for external validation.