Discussion for Plot_assignment.R

I wanted to look at how starvation of the flies during development affects tibia length across the phylogeny, and if the effect of starvation on trait size differed between the sexes. I also wanted to show a visual comparison between the effect of starvation on tibia size to that in wing size.

First, I generated two box plots using untransformed values of tibia length in µm and log2 transformed values of tibia length (log2 µm), respectively, to show the distribution of values for each sex at high condition (HC) and low condition (LC). I was specifically interested in seeing if using the log2 transformed values would reduce the variance in the data caused by very large values for tibia length that were previously noted during data exploration. I suspect that these values are biologically relevant given previous research on exaggerated foreleg size in D. prolongata (See Perdigón Ferreira et al., 2023). Given that I was using values for 30 specimens across 26 species, I thought that a box plot would best represent the distribution of values in my data set, as opposed to displaying the individual points.

To look at the effect of starvation on tibia length in males and females, I used mean tibia length (log2 µm) on my y axis, condition on my x axis, and grouped the data by females (pink) and males (blue). I used facets to generate a graph for each species. While organizing the data into small graphs is not the most effective way to visualize the data according to the Cleveland hierarchy, I thought that including data points for all 26 species on a single graph would make the plot cluttered and therefore harder to interpret. Also, I thought that generating a plot for each species would make for easier comparisons between species. For ease of comparison, I followed this same principle when making a plot for wing size. I used log2 √µm values for wing size to allow for comparison between the two trait, tibia length and wing area. For both graphs, I ordered the levels in my ‘species_full’ variable according to the phylogenetic relationship between the species. The graphs for closely related species were thus displayed beside each other, following the proximity of comparisons principle. The phylogenetic relationship between species was determined using the phylogeny developed in Pesevski 2021 ("DmelSp_phylogeny.png" in my main QMEE repo).

Perdigón Ferreira, J., Rohner, P. T., & Lüpold, S. (2023). Strongly sexually dimorphic forelegs are not more condition-dependent than less dimorphic traits in Drosophila prolongata. Evolutionary Ecology, 37(3), 493-508.

Pesevski, M. (2021). Influence of environmental variation on sexual dimorphism in Drosophila morphology among adaptively diverged populations and in an inter-specific comparative context .