Plasma and urine metabolomic analyses in aortic valve stenosis reveal shared and biofluid-specific changes in metabolite levels

Figures

• 

  Fig 1. Discrimination analysis of AVS patients and control individuals through metabolomic profiling of biofluids. Metabolomic data were derived from GC-MS spectra of urine (a-c) and plasma (d-f) samples from AVS patients (n = 46) and healthy controls (n = 46). Univariate analysis of GC-MS spectral data in the 92 samples was performed to identify metabolomic features significantly separating cases and controls (nominal p<0.05), which are plotted in pink dots in the upper part of the volcano plots (a, d). Multivariate statistics were applied to perform principle component analysis (PCA) (b, e) and orthogonal partial least squares-discriminant analysis (OPLS-DA) (c, f) and assess sample classification in cases and controls. The 95% confidence regions are displayed by shaded ellipses in AVS patients (red) and controls (green).

• 

  Fig 2. Contribution of metabolites in the separation of AVS cases and controls. The Variable Importance in Projection (VIP) was used to weigh the contribution of urine (a) and plasma (b) metabolomic features to the separation between cases and controls in the PLS-DA model. Data were normalized to the internal standard 2-isopropylmalic acid. Upregulated (blue bars) and downregulated (red bars) features are shown. Metabolites found associated with AVS in both urine and plasma are underlined. Details of metabolite features are given in S3 Table.

• 

  Fig 3. Association analysis of urine GC-MS spectral data in AVS patients and control individuals. Data were derived by GC-MS analysis of urine samples from 46 patients and 46 controls. Generalized linear models were used to determine significant associations between metabolomic peaks and AVS after adjusting for age, sex, body mass index, hyperlipidemia and diabetes, and correcting for multiple testing. Signal intensities normalized to the internal standard 2-isopropylmalic acid are plotted against the Q-values. Features showing evidence of statistically significant association with AVS (q values<0.05) are shown with red dots.

• 

  Fig 4. Association analysis of plasma GC-MS spectral data in AVS patients and control individuals. Data were derived by GC-MS analysis of plasma samples from 46 patients and 46 controls. Generalized linear models were used to determine significant associations between metabolomic peaks and AVS after adjusting for age, sex, body mass index, hyperlipidemia and diabetes, and correcting for multiple testing. Signal intensities normalized to the internal standard 2-isopropylmalic acid are plotted against the Q-values. Features showing evidence of statistically significant association with AVS (q values<0.05) are shown with red dots.

• 

  Fig 5. Regulatory pattern of urine metabolites associated with AVS in patients and control individuals. Data from urine candidate metabolites in the 46 AVS cases (orange boxes) and 46 controls (blue boxes) are shown. Data are normalized to the internal standard (2-isopropylmalic acid) and log transformed. The boxplots show the median and the inter-quartile range for each metabolite in the two groups.

• 

  Fig 6. Regulatory pattern of plasma metabolites associated with AVS in patients and control individuals. Data from plasma candidate metabolites in the 46 AVS cases (orange boxes) and 46 controls (blue boxes) are shown. Data are normalized to the internal standard (2-isopropylmalic acid) and log transformed. The boxplots show the median and the inter-quartile range for each metabolite in the two groups.

• 

  Fig 7. Pathway analysis of metabolites associated with AVS. Outputs of urine and plasma metabolomic profiling in AVS patients and control subjects were used to identify changes in biological pathways in the human Kyoto Encyclopedia of Genes and Genomes (KEGG, www.genome.jp/kegg) using the MetaboAnalyst web-tool (www.metaboanalyst.ca). Data are plotted to illustrate the most significantly altered pathways in terms of p-values derived from hypergeometric test on the vertical axis and impact on the horizontal axis.