Metabolite Changes Upon CRC Progres

Metabolite Changes Upon CRC Progression Difffference stage-distributed CRC samples allowed us to investigate the association between metabolite shifts and CRC progression. Based on the 4,526 and 4,765 variables in ESI– and ESI+, using American Joint Committee on Cancer (AJCC) clinical staging, the OPLS-DA analysis of the metabolite profifiles of tumor tissue could separate clusters for each stage (Figure 2). Validation of the OPLS-DA model was performed here by permutation testing. Although the permutation test indicates that the OPLS-DA model is valid, the model fifitting is not very satisfactory (Supplementary Figure 2). There were 94 metabolites exhibiting statistically signifificant difffferential abundance between early- (I, II) and late-stage (III, IV) tumors (VIP > 1.5 and MannWhitney U-test FDR corrected P-value < 0.01), and a total of 48 metabolites were identifified (Supplementary Table 3). Most lipid metabolites showed an increase in late-stage tumors, while dipeptides also showed a decrease in late-stage tumors (Supplementary Figure 3). The results of pathway analysis by Mummichog software indicated that signifificant features are enriched for pathways involved in lipid metabolism (Supplementary Table 4). Metabolite Alterations of CRC Pathologic Characteristics We also sought to determine whether we could identify the difffferences in metabolite features among various histopathological classififications of CRC. The separation of adenocarcinoma and non-adenocarcinoma CRCs was observed using OPLS-DA (Figure 3). Similarly, the permutation test indicates that the OPLS-DA model is valid, but, the model fifitting is also unsatisfactory (Supplementary Figure 4). Forty-three metabolites exhibited statistically signifificant difffferential abundance between adenocarcinoma and non-adenocarcinoma tumors (VIP > 1.5 and P < 0.01). Furthermore, a total of 26 metabolites were identifified (Supplementary Table 5) and almost all these 26 metabolites were lipids. Mummichog indicated that pathways involved in lipid metabolism were also signifificantly enriched (Supplementary Table 5).