Fig. 2: Serum metabolite profiles of frankfurter-fed mice deviate most from control, whereas those fed pork or sausage align more closely.

Serum collected at 56 days underwent metabolomic profiling: a shows scores plots of unsupervised PCA based on the first three principle components (PCs) for Control (n = 9; standard chow (red)), or chow with a 15% incorporation of pork mince (n = 9; blue), pork sausage (n = 10; light blue), or nitrite-containing frankfurter sausage (n = 10; green). Metabolomic data explained 89.4% of the total variation (76.9, 8.0, and 4.5% for PC1, PC2, and PC3, respectively). There was a tendency for profiles from the frankfurter group to separate along the PC1 axis. Profiles from control, pork, and sausages groups clustered more closely with variation more restricted to PC2. b is a Venn diagram showing the number of the 599 metabolites measured which significantly differed in group comparisons with control. c heatmap with hierarchical clustering based on the top 50 metabolites ranked by FDR-corrected P-value (ANOVA). Each row represents a metabolite and each column a sample. Red (high) and blue (low) indicates relative concentrations. Dendrograms indicating similarity are shown at top (samples) on left side (metabolites). Serum metabolite measurements readily clustered mice by diet with pork and nitrite-free sausage being most similar to one another. Frankfurter consumption was characterised by significantly higher levels of certain species of diglycerides (DG), triglycerides (TG), and acylcarnitines (C), and by lower levels of certain phosphatidylcholines (PC/LysoPC) and cholesterol esters (CE). Other notable observations included the differing bile acid responses of taurocholic acid (TCA) and cholic acid (CA). The profound variability of trimethylamine N-oxide (TMAO) responses in frankfurters was also very evident. Control samples are shown in red, minced pork in dark blue, nitrite-free sausage in light blue, and frankfurter (nitrite-containing) in green.