Unified Binding Sites

P04637 Cellular tumor antigen p53 Homo sapiens
UniProtKB AlphaFold PDBe InterPro
93binding sites 544AHoJ-DB entries 5thresholds 7homologs 529holo /264apo

How to read this page

Unified binding site
Each entry's binding site is its set of binding residues, in UniProt numbering. Two sites are compared by Jaccard similarity: the residues they share, divided by the residues of both together. This protein's entries are clustered hierarchically on that similarity (average linkage) and the tree is cut at the threshold; every branch below the cut is one unified site, whose members' sites are on average at least that similar to one another. Each table row is one unified site: Site counts the core residues every member shares (core / union when members differ), Entries the DB entries in it. Hover any cell for the details.
AHoJ-DB entry
One ligand bound to one PDB chain, named pdb-chain-ligand-residue (e.g. 1bab-A-HEM-143: entry 1bab, chain A, heme 143), with the apo and holo structures AHoJ found for that site. This protein's entries are the members of the unified sites and the leaves of the dendrogram.
Threshold
The Jaccard similarity at which the tree is cut. 0.30 gives few, broad sites; 1.00 keeps only identical residue sets together; the page opens on 1.00. LJ is the cut chosen automatically at the largest linkage jump of the dendrogram.
Homologs, interfaces
A unified site can span homologous UniProts (the ×N badge; hover for the list), so its AlphaFold votes may outnumber this protein's entries. ⧉ marks a site at a protein–protein interface.
AlphaFold class
Whether the AlphaFold model of this protein resembles the apo (A) or the holo (H) structures at the site, judged on the pocket; ? = undefined, — = no AlphaFold data. Hover the chip for the pocket- and chain-level breakdown.
Confidence
The probability of the winning class, max(P(apo), P(holo)); the bar fills from white to the class colour.
Density
How close the model sits to the observed structures: 1 = at their peak, 0 = an outlier.
Separation
How distinguishable the apo and holo structures are, 1 − overlap of their distributions: 1 = disjoint, 0 = identical. A classification means little where separation is low; it is blank when one state has no structures.
Apo : Holo
How AlphaFold classifies the site's members (apo, holo). The header strip sums these votes over every site of the threshold the page opens on.
Representative
The exemplar member of a site, shown in indigo in its entries list and in the UBS card; click a cell to pin its card and follow the links.
Binding positions
For each residue of the UniProt sequence, the share (left axis) and number (right axis) of this protein's entries whose binding site includes it; light bars are single-chain sites, blue bars interface sites. The peaks are the protein's recurrent binding sites.
Dendrogram
The clustering tree behind the unified sites. Each leaf is one entry (click to open; the bracket is its number of binding residues), branches join at the similarity on the axis, the dashed lines are the thresholds, and the branches right of a line are that threshold's unified sites, labelled core / union residues.
binding positions · dendrogram

Loading dendrogram for P04637...

Hierarchical clustering (average linkage) of this protein's AHoJ-DB entries by the Jaccard similarity (%) of their binding-residue sets.

  • Leaves: one DB entry each (click one to open it); the number in brackets is its count of binding residues.
  • Dashed lines: the thresholds. The one selected in the table below is highlighted and the others fade; 1.00's lies on the leaves.
  • Branches to the right of a line: that threshold's unified sites, labelled with their core / union residue counts.

Loading binding positions for P04637...

How often each residue of the UniProt sequence is part of a binding site across this protein's entries: the share of entries on the left axis, their count on the right.

  • Light bars: single-chain sites; blue bars: interface sites (two or more chains).
  • Click the plot to enlarge it.
P04637: binding positions
Binding positions of P04637, enlarged
Binding sites at thresholdClustering threshold Jaccard similarity cut on the binding-residue sets of this UniProt's entries, used to merge them into unified sites. Lower = broader (more sites merged); higher = stricter; 1.00 merges only identical residue sets. LJ is a cut picked automatically at the largest jump in the same tree (UniProts with ≥3 entries only).:
Loading binding sites…
Loading binding sites…
Loading binding sites…
Threshold: 1.0
Cluster AlphaFold model (unified site)
UBS Unified Binding Site Cluster of this UniProt's binding sites merged at this threshold (Jaccard similarity of residue sets). ⧉ marks a binding site at a protein–protein interface. Site Site size (residues) Number of core residues shared by all cluster members. A second number (core / union) appears when the union across members is larger. Hover a cell for the residue list. Entries AHoJ-DB entries Structures assigned to this cluster. Hover for the list — each links to its DB entry. Class AlphaFold classification Predicted state (H holo, A apo, ? undefined, — no AlphaFold data) of the unified binding site — this site merged across its homologs, which may span several UniProts. Hover for P(apo)/P(holo), votes, density, separation and the chain-level values. Confidence Classification confidence Probability of the winning AF class for the unified site, max(P(apo), P(holo)), averaged over the pocket metrics and the site's members. The bar fills from white to the class colour (blue apo, red holo, as in the violin plots); — when the site has no AlphaFold probabilities. Density Normalised density Density of the observed apo + holo structures at the AlphaFold model's position, averaged over the pocket metrics and the site's members: 1 = the model sits at the peak of the observed distribution, 0 = it is an outlier. Hover a cell for the chain-level value. Separation Apo / holo separation How distinguishable the apo and holo distributions are: 1 − overlap, averaged over the overall and the mode (peak) overlap, the pocket metrics and the site's members. 1 = disjoint, 0 = identical, so a classification means little where this is low. Hover a cell for the chain-level value. Apo:Holo AF apo : holo votes How the unified site's members are classified by AlphaFold (blue apo, red holo, as in the violin plots). Votes span all homologous members across UniProts, so the total can exceed this row's Entries count.
#1⧉ 6 4 H0.72
0.72
 0.19 0.76
1:3 A:H
#2⧉ 5 9 H0.85
0.85
 0.27 0.59
1:8 A:H
#3 5 15 H0.87
0.87
 0.38 0.40
0:15 A:H
#4 7 2 H0.70
0.70
 0.68 0.35
0:2 A:H
#5 6 9 H0.59
0.59
 0.61 0.43
0:9 A:H
#6⧉ 6 2 H0.59
0.59
 0.47 0.32
0:2 A:H
#7⧉ 5 1 H0.80
0.80
 0.34 0.34
0:1 A:H
#8 6 5 H0.68
0.68
 0.71 0.39
0:5 A:H
#9 5 261 H0.66
0.66
 0.67 0.39
1:260 A:H
#10⧉×2 6 1 — — — — —
#11 6 1 H0.57
0.57
 0.81 0.45
0:1 A:H
#12 6 1 H0.67
0.67
 0.76 0.41
0:1 A:H
#13 4 52 H0.76
0.76
 0.67 0.38
0:52 A:H
#14⧉ 5 1 H0.69
0.69
 0.31 0.39
0:1 A:H
#15 15 1 H0.74
0.74
 0.83 0.47
0:1 A:H
#16 4 2 A0.59
0.59
 0.57 0.20
2:0 A:H
#17 5 7 A0.63
0.63
 0.68 0.23
7:0 A:H
#18 4 1 A0.67
0.67
 0.56 0.17
1:0 A:H
#19 3 1 H0.53
0.53
 0.76 0.09
0:1 A:H
#20⧉ 9 1 A0.62
0.62
 0.90 0.50
1:0 A:H
#21 9 1 A0.54
0.54
 0.74 0.12
1:0 A:H
#22 10 1 A0.56
0.56
 0.72 0.12
1:0 A:H
#23 10 1 A0.54
0.54
 0.61 0.11
1:0 A:H
#24 9 1 A0.54
0.54
 0.66 0.10
1:0 A:H
#25 9 1 A0.53
0.53
 0.75 0.10
1:0 A:H
#26 10 1 A0.54
0.54
 0.68 0.15
1:0 A:H
#27 12 1 H0.50
0.50
 0.54 0.11
0:1 A:H
#28 8 1 A0.57
0.57
 0.57 0.18
1:0 A:H
#29⧉ 8 1 H1.00
1.00
 0.12 —
0:1 A:H
#30⧉ 7 1 H1.00
1.00
 0.15 —
0:1 A:H
#31 5 1 H0.64
0.64
 0.85 0.30
0:1 A:H
#32 5 1 A0.56
0.56
 0.73 0.14
1:0 A:H
#33 7 2 A0.56
0.56
 0.40 0.14
2:0 A:H
#34 7 1 A0.53
0.53
 0.44 0.11
1:0 A:H
#35 8 1 A0.56
0.56
 0.47 0.14
1:0 A:H
#36 6 5 A0.54
0.54
 0.57 0.17
5:0 A:H
#37 6 3 A0.63
0.63
 0.56 0.17
3:0 A:H
#38 5 6 A0.61
0.61
 0.51 0.16
6:0 A:H
#39 6 1 A0.53
0.53
 0.70 0.18
1:0 A:H
#40 5 3 A0.56
0.56
 0.82 0.15
3:0 A:H
#41 6 1 A0.52
0.52
 0.92 0.09
1:0 A:H
#42 6 2 A0.57
0.57
 0.66 0.12
2:0 A:H
#43 7 2 A0.55
0.55
 0.61 0.14
2:0 A:H
#44 8 1 A0.54
0.54
 0.66 0.10
1:0 A:H
#45 6 3 A0.68
0.68
 0.56 0.24
3:0 A:H
#46 7 3 A0.65
0.65
 0.55 0.18
3:0 A:H
#47 5 1 A0.61
0.61
 0.46 0.31
1:0 A:H
#48 7 2 A0.69
0.69
 0.61 0.58
2:0 A:H
#49 9 3 A0.65
0.65
 0.54 0.57
3:0 A:H
#50 16 2 A0.81
0.81
 0.83 0.77
2:0 A:H
#51 17 2 A0.74
0.74
 0.82 0.51
2:0 A:H
#52 16 1 A0.71
0.71
 0.92 0.52
1:0 A:H
#53 18 2 A0.72
0.72
 0.81 0.54
2:0 A:H
#54 17 2 A0.70
0.70
 0.83 0.51
2:0 A:H
#55 15 2 A0.74
0.74
 0.80 0.67
2:0 A:H
#56 16 7 A0.70
0.70
 0.82 0.53
7:0 A:H
#57 17 1 A0.72
0.72
 0.78 0.53
1:0 A:H
#58 12 2 A0.72
0.72
 0.88 0.59
2:0 A:H
#59 12 1 A0.76
0.76
 0.81 0.55
1:0 A:H
#60 12 4 A0.74
0.74
 0.85 0.56
4:0 A:H
#61 13 1 A0.78
0.78
 0.75 0.55
1:0 A:H
#62 15 5 A0.76
0.76
 0.83 0.59
5:0 A:H
#63 14 1 A0.79
0.79
 0.77 0.56
1:0 A:H
#64 13 2 A0.73
0.73
 0.83 0.55
2:0 A:H
#65 14 4 A0.77
0.77
 0.87 0.58
4:0 A:H
#66 16 2 A0.77
0.77
 0.87 0.61
2:0 A:H
#67 17 2 A0.77
0.77
 0.80 0.57
2:0 A:H
#68 14 12 A0.74
0.74
 0.84 0.55
12:0 A:H
#69⧉ 15 1 H0.61
0.61
 0.43 0.18
0:1 A:H
#70⧉ 14 1 A0.55
0.55
 0.57 0.75
1:0 A:H
#71 16 5 A0.75
0.75
 0.70 0.54
5:0 A:H
#72 15 5 A0.74
0.74
 0.75 0.55
5:0 A:H
#73 14 12 A0.65
0.65
 0.78 0.53
12:0 A:H
#74 16 2 A0.68
0.68
 0.73 0.51
2:0 A:H
#75 14 1 A0.60
0.60
 0.75 0.51
1:0 A:H
#76 10 1 A0.74
0.74
 0.72 0.69
1:0 A:H
#77 12 1 A0.77
0.77
 0.85 0.69
1:0 A:H
#78 14 1 A0.64
0.64
 0.70 0.48
1:0 A:H
#79 13 1 A0.71
0.71
 0.84 0.43
1:0 A:H
#80 11 1 A0.68
0.68
 0.73 0.53
1:0 A:H
#81 12 1 A0.67
0.67
 0.63 0.49
1:0 A:H
#82 15 1 A0.69
0.69
 0.94 0.41
1:0 A:H
#83⧉ 4 1 A0.54
0.54
 0.60 0.36
1:0 A:H
#84 3 1 A0.52
0.52
 0.56 0.36
1:0 A:H
#85 6 1 — — — — —
#86⧉×7 1 3 H0.50
0.50
 0.13 0.19
106:81 A:H
#87⧉×3 1 13 H0.91
0.91
 0.62 0.45
3:55 A:H
#88⧉×4 1 2 H0.74
0.74
 0.42 0.86
1:18 A:H
#89⧉×2 1 1 H1.00
1.00
 0.65 —
0:6 A:H
#90⧉×2 1 1 H1.00
1.00
 0.25 —
0:1 A:H
#91⧉×2 1 1 A0.58
0.58
 0.23 0.35
1:0 A:H
#92⧉×2 1 1 H0.52
0.52
 0.19 0.79
0:1 A:H
#93 4 1 A0.54
0.54
 0.77 0.12
1:0 A:H
Loading binding sites…