Fig 1.
Experimental design and study workflow.
Outline the workflow followed in this study. Briefly, plasma and stool samples of wild type and Tg-McGill animals were collected, and metabolomic analysis and 16sRNA profiling were performed to obtain discriminant metabolites through sPLS-DA. Metabolite-protein associations were extracted from public databases. First-order protein-protein interactions (PPI) were generated and networks with a large overlap were merged, producing a final set of 657 modules. “Core modules” (n = 20) were defined through over-representation analysis. Finally, analysis of distinct biological signals was performed to determine the most relevant ranked networks.
Fig 2.
Discriminant metabolites in each experimental group.
(A) Diagnostic plot from multiblock sPLS-DA. Samples are represented based on the specified component (here ncomp = 2) for each dataset (microbiota_clr, metabolon_feces and metabolon_plasma) and colored by experimental group. The 95% confidence ellipse plots are represented. The bottom left numbers indicate the correlation coefficients between the first components from each dataset. (B) Sample plot from sPLS-DA. Samples are projected into the space spanned by the first two components showing discrimination between the experimental groups. Numbers in box represent the Classification Error Rate (Overall.ER) and Barret Error Rate (Overall.BER) scores. (C) Clustered Image Map (CIM) for the variables selected by best multiblock sPLS-DA performed on the study on component 1 and 2. The CIM represents samples in rows (indicated by their experimental group on the left side of the plot) and selected features in columns (indicated by their data type at the top of the plot).
Table 1.
AD-core networks.
Table 2.
Exploratory networks.
Table 3.
Metabolite Heatmap.
Fig 3.
Bottom row, Top 20 exploratory Networks IDs. Left, significantly overrepresented Disease Ontology (DO) and manually collected OMIM categories related to AD. DO-OMIM class significance and enrichment score are calculated as the geometric mean of the p-values and mean of the enrichment scores, respectively, for all contained DO-OMIM categories.
Fig 4.
Bottom row, Top 20 exploratory Networks IDs. Left, significantly overrepresented Reactome and WikiPathways categories related to AD manually collected. Reactome and WikiPathways class significance and enrichment score are calculated as the geometric mean of the p-values and mean of the enrichment scores, respectively, for all contained Reactome and WikiPathways categories.
Fig 5.
Bottom row, Top 20 exploratory Networks IDs. Left, significantly overrepresented GO categories related to AD manually collected. GO class significance and enrichment score are calculated as the geometric mean of the p-values and mean of the enrichment scores, respectively, for all contained GO categories.
Fig 6.
Structural features of the C1_3_65 and C1_3_70 networks.
Circle, gene/protein annotated. Line, gene/protein with bibliographic evidence of association. First column, gene/protein labeled with pathway association; Second column, gene/protein labeled with AD gene sets association; Third column, gene/protein labeled with differential metabolite association.