Figure 4
Principal component analysis (PCA) of the set of qrr reporter fluorescence histograms. Under each of 49 combinations of C8HSL and 3OC6HSL concentrations, the fluorescence measurements for ~ 400 individual cells were histogrammed into 20 bins and a square root transformation was applied to the bin heights (Methods). The PCA of the complete set of histograms then provides (a) scores for each principal component as a function of HSL input condition, and (b) loadings associated with each PCA component. The loadings describe the fluorescence-dependence described by each PCA component, and the scores describe the HSL dependence of that component. The percentage of data variance explained by each PCA component is indicated on the panel. The first four PCA components only are shown; these together account for 94.5% of the variance in the dataset (see text). HSL axes in (a) switch from linear to logarithmic at 10 concentration units48.
