Principal component analysis (PCA) of variation across the entire data set.

<p><b>A</b> Pareto plot of the percentage of the variance explained by each principal component (black bars), and the cumulative sum of the explained variance (dark grey line). <b>B</b> Empirical cumulative distribution function in PC space for each class and for principal components one to four. <b>C</b> Contribution of each cellular parameter to the variation of the first four principal components. See main text for an interpretation of these results.