blob: a06430b6fad186036527d1f49aa1dfc0e62da9f9 (
plain)
1
|
<p>High-fat diets are associated with increased obesity and metabolic disease in mice and humans. Here we used analysis of variance (ANOVA) to scrutinize a microarray data set consisting of 10 inbred strains of mice from both sexes fed atherogenic high-fat and control chow diets. An overall F-test was applied to the 40 unique groups of strain-diet-sex to identify 15,288 genes with altered transcription. Bootstrapping k-means clustering separated these changes into four strain-dependent expression patterns, including two sex-related profiles and two diet-related profiles. Sex-induced effects correspond to secretion (males) or fat and energy metabolism (females), whereas diet-induced changes relate to neurological processes (chow) or immune response (high-fat). The full set of pairwise contrasts for differences between strains within sex (90 different statistical tests) uncovered 32,379 total changes. These differences were unevenly distributed across strains and between sexes, indicating that strain-specific responses to high-fat diet differ between sexes. Correlations between expression levels and 8 obesity-related traits identified 5,274 associations between transcript abundance and measured phenotypic endpoints. From this number, 2,678 genes are positively correlated with total cholesterol levels and associate with immune-related categories while 2,596 genes are negatively correlated with cholesterol and connect to cholesterol synthesis. Keywords: gene expression analysis, strain comparision, effect of dietary fat, sex-specific effects</p>
|
