Fig. 4: haCCA generated multi-modal data allows integrative analysis of transcriptome and metabolome.

A UMAP visualization of clustering results from three data modalities: metabolic clusters from MALDI-MSI (left), integrated clusters after \({haCCA}\) analysis (middle), and transcript clusters from spatial transcriptomics (right), revealing distinct spatial organization patterns. Cluster labels were used for visualization. B Sankey diagram depicting the relationships between transcript clusters (S0–S6), integrated clusters (C0–C7), and metabolic clusters (M1–M5). Width of connecting lines represents the proportion of shared spots between clusters. C Spatial visualization of cluster distributions demonstrating the preservation of both metabolic and transcriptomic features. Left to right: metabolic clusters, integrated clusters after \({haCCA}\), and transcript clusters. Red boxes highlight regions where metabolic features are predominantly preserved; green box highlight regions where transcript features are predominantly preserved. D Dot plot showing marker metabolites for integrated clusters. Dot size indicates fraction of spots expressing each metabolite, while color intensity represents mean expression level. E Dot plot showing marker gene expression across integrated clusters. Dot size represents fraction of expressing spots, and color intensity indicates mean expression level.