Fig 1.
Depiction of proteome sorting on the basis of maximum local composition.
(A) For each amino acid and window size combination, each yeast protein is sorted into percent composition bins based on the maximum local composition of the amino acid within the given sliding window size. This effectively sorts the yeast proteome 200 distinct ways (20 amino acids x 10 different sliding window sizes). (B) Visual representation of proteins sorted based on maximum local aspartic acid composition with a 20 amino acid sliding window.
Fig 2.
Distribution of the yeast proteome based on maximum local amino acid composition.
The number of proteins partitioned into each window size/percent composition bin for each of the 20 canonical amino acids are plotted as a function of maximum local composition for each window size. Scatter points are connected by line segments for visual clarity only.
Fig 3.
Maximum local amino acid composition is associated with residue-specific differences in protein abundance.
For each amino acid, protein abundance values corresponding to proteins partitioned into a given window size and percent composition bin were compared to values for all proteins of length ≥ the corresponding window size that were excluded from the bin. For this figure and related subsequent figures, red and blue points indicate bins for which the distribution of protein abundance values differ significantly (Bonferroni-corrected p ≤ 0.05) from those of excluded proteins: red points indicate bins with a lower median value relative to that of excluded proteins, whereas blue points indicate bins with a higher relative median value. Grey points indicate comparisons lacking statistical significance. Individual points are scaled within each subplot to reflect the sample sizes of proteins contained within each bin.
Fig 4.
Maximum local amino acid composition is associated with residue-specific differences in protein half-life.
For each amino acid, protein half-life values corresponding to proteins partitioned into a given window size and percent composition bin were compared to half-life values for all proteins of length ≥ the corresponding window size that were excluded from the bin.
Fig 5.
Maximum local amino acid composition is associated with residue-specific differences in nTE.
For each amino acid, nTE values corresponding to proteins partitioned into a given window size and percent composition bin were compared to nTE values for all proteins of length ≥ the corresponding window size that were excluded from the bin.
Table 1.
The life-cycle of proteins with high local composition of individual amino acids involves the coordinated regulation of translation efficiency, protein abundance, and protein half-life.
For each amino acid, trends in median values for nTE, protein abundance, and half-life upon enrichment (i.e. approaching higher percent compositions) of the given amino acid are indicated. “Higher” indicates that proteins in larger percent composition bins tend to have a larger median value compared to all other proteins, while “Lower” indicates that proteins in larger percent composition bins tend to have a larger median value compared to all other proteins. “Mixed” indicates amino acids for which multiple transitions are observed upon progressive compositional enrichment. Datasets without clear, statistically significant transition thresholds are also indicated (“n.s.”). Colors correspond to the colors used in Figs 3–5.
Fig 6.
Relationships between maximum local composition and protein abundance in C. elegans.
For each amino acid, protein abundance values corresponding to proteins partitioned into a given window size and percent composition bin were compared to values for all proteins of length ≥ the corresponding window size that were excluded from the bin. Trends in protein abundance as a function of maximum local composition for many amino acids are remarkably similar between yeast and C. elegans (compare with Fig 3).
Fig 7.
Maximum local amino acid composition corresponds to protein-protein interaction promiscuity in a residue-specific manner.
Local enrichment for individual amino acids corresponds to composition-dependent changes in the number of unique protein-protein interaction partners.
Fig 8.
Cell model depicting predominant functions of CEDs.
Residue-specific CEDs are associated with both overlapping and distinct functions. Main cellular/molecular processes associated for each type of CED are derived from significantly enriched (Bonferroni-corrected p ≤ 0.05) GO-terms in S3 Table.
Fig 9.
Compositional enrichment profiles associated with major subcellular compartments.
All plotted points indicate protein sets for which association with the indicated subcellular compartment is statistically significant (Fisher’s exact test, with Bonferonni-corrected p < 0.05). Warm colors (reds, oranges, and yellows) correspond to charged residues. Green colors indicate polar residues. Cool colors (purples and blues) correspond to hydrophobic and aromatic residues respectively.
Fig 10.
Composition profiles associated with membraneless organelles.
All colored points indicate minimum percent composition thresholds for which components of stress granules (A) or P-bodies (B) are significantly enriched (p < 0.05). Only amino acids for which significant enrichment of stress granule or P-body proteins was observed in at least two composition bins are shown. For greater sensitivity, plots were generated using uncorrected p-values. Therefore, any individual point should be viewed with some skepticism: however, the presence of multiple consecutive significant points within each window size suggest that the observed trend is likely not an artifact of multiple hypothesis testing.