Due: Fri June 6
For each question provide: (i) a written description of the results of your analyses including a brief justification for the test/analysis you used, (ii) the R source file used to produce these results, (iii) a clear visualization of both the underlying data points and the variability in that data provided to you and a description of what is being visualized (e.g. like in a figure legend).
Q1: You want to test the hypothesis that a conditional knockout of a gene of interest in the hypothalamus affects weight in mice in regular diet (RC) and high-fat diet (HFD). You perform independent knockout experiments for ten mice, randomly place five knockout and wild-type mice on RC and HFD and measure the weight of each mouse. The data is provided here.
Is there evidence for a difference in the weight of the knockout and wild-type mice, and if so, is this difference dependent on the diet the mice are given? Use two-way ANOVA with an interaction term to evaluate whether the effect of the gene knockout on weight depends on diet. You can also use the R function ‘interaction.plot()’ to visualize a potential interaction.
What would be your conclusion from these results? Are there any potential caveats to these analyses that might make you question your results and if so how might you design the study differently?
(10 points)
Q2: You want to test the hypothesis that activated microglia isolated from the substantia nigra of familial Parkinson’s disease (PD), sporadic PD and control samples obtained from multiple biorepositories have differences in pro-inflammatory cytokine secretion. Previous studies have shown that there is large variability in sample preparation and quality and experimental measurements across different biorepositories, and therefore you want to include biorepository as a blocking factor in your analysis. You perform your experiments for PD and control samples from each biorepository and measure levels of the cytokines IL-1a, TNFa and C1q, and the data are stored here.
Is there evidence for a difference in levels of any of the three cytokines from activated microglia between disease groups? Use two-way ANOVA with biorepository as a blocking factor to determine whether cytokine levels differ across disease group. Make sure to check the distribution of the values, and if non-normal find an appropriate transformation for the data prior to running ANOVA. When plotting, make sure to separate the values based on the blocking factor - for example based on facets, or different coloring.
(10 points)