Fig. 2: Validation of tissue-specific GEMs through regional metabolomics and pathway analysis.

a Schematics showing the metabolomics analysis performed on four Drosophila regions, followed by KEGG over-representation analysis. Elements created in BioRender. Moon, S. (2026) https://BioRender.com/1nasxv2. b Principal component analysis (PCA) of metabolite profiles across regions (n = 4). c Heatmap showing the metabolite profiles across thorax, head, abdomen, and gut. d Bubble plot showing enriched metabolites in each region compared to others. One-way ANOVA followed by Benjamini–Hochberg (BH) correction was used for multiple comparisons. e Over-representation analysis of enriched metabolite sets across regions, using a one-sided hypergeometric test, followed by BH correction for multiple comparisons. f Heatmap of weighted Jaccard index values showing pathway overlaps between metabolomics-derived and GEM-predicted pathways. g, h Boxplots showing weighted Jaccard index distributions between metabolomics-derived pathways and either g Muscle-GEM or h Fat body-GEM. Center line is median; box limits are first and third quartiles; whiskers are 1.5 × interquartile range; red points are mean; gray points present individual bootstrap samples (n = 10,000), shown in full including those outside the whiskers. Statistical significance is based on comparisons against the reference region (g: thorax; h: abdomen) using one-way ANOVA followed by Bonferroni-corrected pairwise tests; *** p < 0.001. Overlapping pathways between GEM-predicted and region-specific enriched pathways are shown to the right. Source data are provided as a Source Data file.