Table 1.
Post translational modification datasets.
Figure 1.
Selectivity of post-translational modification on human protein complexes.
Median modification level of protein complexes plotted against total number of modifications for each PTM, with each data point representing a unique complex for (A) acetylation, (B) tyrosine phosphorylation, (C) ubiquitination and (D) serine/threonine phosphorylation. Data points may overlay preventing visualisation of each unique complex. The density data from 100 random datasets is overlaid on each plot with graded colours representing percentages of total randomized data. Complexes that are highly unlikely to be generated through random dataset generation are selected at confidence cut-offs of 99% for Ac, pY and Ub and 95% for pS/T and highlighted in red. (E) Overlap analysis for selected, highly modified complexes from 1A–D. (F) GO analysis highlighting coordinated differential molecular function control by each group of complexes selected in the above analysis. Ac & Ub represents the 57 complexes that are enriched for both these PTMs, ≥3PTMs represents the 39 complexes enriched in at least 3 PTMs.
Figure 2.
Signal integration on highly modified protein complexes.
(A) Percentage of total modifications across each specific group of complexes in comparison to all complexes in the dataset and all of the enriched complexes combined (Selected Complexes). Box plot represents the total number of modifications present across all complexes for each enriched group. (B) Schematic representations of potential PTM integration. Putative signal integration through distinct singly-modified subunits (i) is compared to more complex combinatorial signaling through multiply-modified subunits (ii). (C) Enrichment analysis for each subgroup of nodes. Each square represents the Z-score for an increase or decrease in signal compared to random samples from the entire complex dataset. * Represents Z-scores >35, not reflected by the colour code of the heatmap.
Figure 3.
PTMi spot identification and characterisation.
(A) 2D density plot of local STYK AA density windows plotted against local PTM density with a histogram showing the number of local peaks. The 500 most outlying data points are plotted as points on the density plot, however with substantial overlay, preventing visualization of all data point. Some PTM regions contain a higher PTM density than modifiable residues, highlighting regions which are key candidates for direct ubiquitination:acetylation competition on the same lysine residues. Specific proteins with densely modified regions are indicated through coloured circles. (B) 2D density plot of the number of 20 AA windows containing more than 1 PTM. (C) The number of high density windows that contain multiple PTMs compared to 100 random annotation permutation simulations. (D) Overlap analysis between individual PTM 20AA windows and annotated protein domains. Int: 20AA window internal to a protein domain, Lg: Large overlap with a proteins domain (>10AAs), Sm: Small overlap with a protein domain, Ext: 20AA window external to an annotated protein domain. (E) Overlap analysis between individual PTM 20AA windows and predicted protein disorder. High: Every amino acid is predicted to be disordered, Med: 11–19 AAs in a window are predicted to be disordered, Low: 1–10 AAs in a window are predicted to be disordered. (F) Frequency of 20AA windows across a protein sequence that are mutated in cancerous cells, sorted based on their protein domain annotation (coloured bars). The frequency of 20AA windows outside of annotated protein domains that are mutated in cancerous cells, sorted based on their PTM density (Grey bars).
Figure 4.
PTMi spot containing proteins.
(A) Breakdown of the PTMs present in each of the 405 PTMi spots identified here. (B) Examples of single phospho-PTMi spot containing proteins with histograms of the local PTM density across the protein sequence. Colour codes beneath each density histogram represent the number of distinct PTMs (red scale) and the local STKY density (yellow∶blue scale, see legend [is in C]). (C) Multi-signal PTMi spot proteins. Examples of proteins annotated either as canonical PTMi spot containing protein, a protein that integrates both ubiquitin and multiple phosphorylations in a PTMi spot (putative degron), or to one of five key regulatory cellular modules. Representative examples of PTM density distributions of a protein in each sub-group is linked to the panel, a full list of PTMi spot containing proteins is found in Dataset S4.