Fig. 3.

Mechanosensitive metabolites in mouse synovial fluid. (A) PLS-DA VIP scores determined the top 25 mechanosensitive metabolite features plotted by their m/z value that contributed the most to the variation between datasets. The majority of metabolite features with high VIP scores were higher in exercised than controls. These features are matched to metabolite identities in Table 1. Select mechanosensitive metabolite features higher in exercised than control SF matched with metabolite identities N-acetyl tryptophan and phosphatidylserine. (B) Cumulative distribution of pointwise biserial correlations for metabolite features correlated to acute exercise. Features with a correlation coefficient greater than 0.7 or less than −0.7 were considered accumulated (blue) or depleted (red), respectively. The top 10 accumulated and top 10 depleted metabolite features were considered mechanosensitive metabolites and matched with metabolite identities in Table 2 (C) Volcano plot analysis shows metabolite features significantly upregulated or downregulated by exercise. Metabolite features in the upper right and left regions (dashed lines) have a p-value<0.05 and greater than twofold change. The significant metabolite features are matched to metabolite identities in Table 3 (D) Venn Diagram showing metabolites used in fold-change analysis. There was substantial overlap of metabolites from the volcano plot and VIP analysis. However, there was limited overlap between the metabolites used from the correlation analysis. Note m/z ratios (mass-to-charge) may match multiple metabolite identities depending on adduct formation (H+ or Na+) occurring during ionization.