Minor revisions to TTR and respirometry analysis

We’re going to pretend that I’ve actually thought about this since last December…YIKES. Anyways, today I revisited my code to ensure I actually accounted for demographic variables.

Turns out I did account for demographic differences in my TTR code! I tossed all variables in my model and then used a backwards deletion approach to pick the best fit model. The best model for TTR included day and weight even though they were non-significant, so I just made sure that the model without those variables didn’t have a lower AIC. I also used emmeans to conduct post-hoc tests for temperature. I guess they aren’t strictly necessary for temperature since the model compares 13ºC vs. either 30ºC or 5ºC, but it did give me a value to compare 30ºC vs. 5ºC. The output from the post-hoc test can be found here. Based on the statistical output, I made some slight changes to the statistical information in the figure. Initially I thought I could add some letters on top of the boxplot to distinguish differences, but because there was no treatment x day interaction, the visualization may be a bit harder to digest. So we’re sticking with text in the white space.

Figure 1. TTR figure

I also accounted for demographic differences in my respirometry analysis! The most parsimonious model was one that included treatment and day, but not the interaction nor any other demographic variables. I also conducted posthoc tests, and saved the output showing differences by treatment and day. Again, adding letters to the figure for these post-hoc would be more confusing since there wasn’t an interaction, so I settled for updating the stats on the figure.

Figure 2. Respirometry figure

And that’s that! Next up: adding PERMDISP to my metabolomics and lipidomics analyses.

Going forward

1. Add PERMDISP to clarify effects are due to centroid variance and not spread
2. Integrate physiology and -omics data with WCNA
3. Metabolomics and lipidomics DIABLO analysis
4. Outline discussion
5. Write discussion
6. Write abstract