RNA-seq analysis of 6-month-old zebrafish models of EOFAD. (A) Schematic of an RNA-seq experiment using zebrafish. A single mating of a single pair of fish heterozygous for either an EOFAD-like or a non-EOFAD-like mutation results in a family heterozygous mutant, transheterozygous mutant and wild-type siblings. Comparisons made between genotypes in an RNA-seq experiment are depicted. (B) Heatmap summary of significantly altered KEGG and IRE gene sets in zebrafish EOFAD genetic models at 6 months of age. Only gene sets significantly altered (FDR-adjusted harmonic mean P<0.05) in at least two comparisons of mutant zebrafish to their corresponding wild-type siblings are shown. Columns are grouped by whether or not the zebrafish genotype is EOFAD-like, and rows are clustered based on their Euclidean distance. The numbers are FDR-adjusted harmonic mean P-values. (C,D) Heatmap indicating the log fold change (logFC) of genes in the KEGG gene sets for oxidative phosphorylation (C) and the ribosome (D) in zebrafish mutants compared to their wild-type siblings. Rows are clustered based on their Euclidean distance, and columns are grouped by the complex in the electron transport chain to which an encoded protein belongs (C), or whether an encoded protein forms part of the large or small ribosomal subunits (D). Only genes considered detectable in all RNA-seq experiments are depicted. See Figs S1-Table 1 for more information on individual study designs.

Microarray analysis of male 12-month-old homozygous App^NL-G-F mice. (A) Principal component analysis of brain transcriptome data from male 12-month-old homozygous App^NL-G-F (n=3), 3xTg (n=3), App wild-type (App^+/+, n=3) and non-transgenic (non-Tg, n=3) mice. The numbers in parentheses indicate the percentage of variation in the dataset explained by a principal component. Each point represents a sample, which are coloured by genotype. (B) Bar chart showing the FDR-adjusted P-value (directional hypothesis) from fry on marker genes of neurons, oligodendrocytes, astrocytes and microglia in App^NL-G-F relative to wild type. (C) Heatmaps indicating the expression (intensity) of genes within these marker gene sets summarised using K-means (K=4). (D) Heatmap showing the expression of genes in the KEGG_LYSOSOME gene set, clustered by their Euclidean distance. Each gene is labelled in red if they were identified as differentially expressed (DE), and the magnitude of the fold change (logFC) is shown in green.

RNA-seq analysis of 3-month-old APOE-TR mice. (A) Visual representation of the comparison of APOE4 (n=7 males and 9 females) or APOE2 (n=8 males and 8 females) mice to APOE3 (n=8 males and 8 females). This comparison was made for both male and female mice separately. (B) PCA of 3-month-old APOE-TR mice. Principal component 1 (PC1) is plotted against PC2. The numbers in parentheses indicate the percentage of variation in the dataset explained by a principal component. In the left graph, each point represents a sample, which are coloured by APOE genotype and shaped by sex. In the right plot, each point is coloured according to litter (implied from the date of birth of each mouse), and shaped by APOE genotype. (C,D) Heatmap showing the log fold change (logFC) of genes in the KEGG_OXIDATIVE_PHOSPHORYLATION (C) and KEGG_RIBOSOME (D) gene sets in APOE-TR mice. Rows are clustered based on their Euclidean distance, and columns are grouped by the complex in the electron transport chain to which an encoded protein belongs (C), or whether an encoded protein forms part of the large or small ribosomal subunits (D). Genes are labelled in blue above if they were classified as differentially expressed (DE, FDR<0.05) in the differential gene expression analysis in the listed comparisons.