Figure 2

Experimental workflow and distribution of protein aggregation phenotypes. (A) Workflow of high-throughput protein solubility measurement and classification. (B) Distribution of cellular fluorescence phenotypes based on 1,000 classified cells for each overexpressed cytoplasmic protein (represented by dots). The location of each dot is calculated from the fraction of cells showing each phenotype, with red lines representing the decision boundaries between aggregation categories assigned. The majority of proteins are located close to the vertices, demonstrating that cells typically show homogenous aggregation behavior. (C) Comparison of in vivo and in vitro solubility phenotype of proteins. In vitro aggregating proteins show a strong overlap with proteins forming either dark or fluorescent aggregates in vivo (odds ratio = 14.5, P < 10^–10, Fisher’s exact test). (D) Frequency of proteins according to their aggregation phenotypes.