Linking mass spectrometry data to tumour activity: A preliminary investigation

Objective To identify which metabolites and pathways distinguish tumour from non-tumour regions in patient tissues and explain their relevance using established cancer hallmarks. Background Tumours rewire cellular metabolism, leaving detectable fingerprints in mass spectrometry (MS) profiles. Biological pathways involve coordinated changes across many metabolites; hence, patterns across thousands of m/z values were condensed into the importance spectrum because single peaks rarely tell the full story of cellular activity. Methods I developed and utilised an interpretation pipeline that connects mass-to-charge (m/z) feature portraits to metabolites, maps those metabolites to pathways, and relates pathway activity to established cancer hallmarks. Discriminative m/z peaks (retaining features ≥ 98 m/z) were identified per sample and ranked in an importance spectrum. This spectrum is a profile indicating each feature’s contribution to distinguishing tumour vs non-tumour in tissue. Putative metabolite identities were identified via the Human Metabolome Database (HMDB), then aggregated to biological pathways using the Small Molecule Pathway Database (SMPDB). The results in a pathway-level view that contextualises isolated peaks as parts of coordinated biochemical processes. To ground these signals in established biology, identified pathways were cross-referenced to canonical cancer hallmarks. Each pathway was correlated to one or more hallmarks with support from literature. This provides a principled bridge from ion signals to recognised cancer biology. Results In patient tissues, high-importance m/z features are frequently mapped to upregulated metabolites in dysregulated pathways affected by tumours. This is consistent with metabolic reprogramming seen in cancer biology. Conclusion In conclusion, the metabolite and pathway patterns correlate with well-defined cancer behaviours that boost tumourigenesis, metastasis, and a pro-angiogenic microenvironment.