Fig. 4

Unsupervised clustering of fractionation mass spectrometry data. a. Principal component plot of components 1 and 2 calculated from standardized average protein intensities. Standardized protein intensities across fractionation samples were clustered using SOM (self-organizing maps). Proteins in PCA space are colored according to their assignment to one of eight distinct clusters (see color code in b). 795 proteins could not be associated with any of the clusters (gray colored dots). b. Line plots of the median standardized protein intensities across fractionation steps of the different clusters assigned by SOM (see a). c. Heatmap summarizing enrichment analyses of cellular localization annotations in the respective clusters. Color darkness corresponds to levels of enrichment (−log10 of adjusted P-values). Rectangles and circles indicate highest enrichment for the corresponding localization and gene clusters, respectively. General and more specific localization categories are framed in black and red, respectively