Fig. 1.

Overview of the data analysis strategy. The TEDDY case-control dataset was used. Preprocessing entailed data binning, missing value imputation, and within case-control normalization. Statistical analysis proceeded through NPLS-DA models on islet autoimmunity for each omics modality, selection of variables using the VIP statistics, and a final multi-omics NPLS-DA model with selected variables. Integrative biological interpretation used enrichment analysis to identify relevant pathways, partial correlation analysis to find novel gene/metabolite relationships, and PaintOmics3 to add pathway models to multi-omics data visualization