Structural templates for comparative protein docking

Abstract

Structural characterization of protein-protein interactions is important for understanding life processes. Because of the inherent limitations of experimental techniques, such characterization requires computational approaches. Along with the traditional protein-protein docking (free search for a match between two proteins), comparative (template-based) modeling of protein-protein complexes has been gaining popularity. Its development puts an emphasis on full and partial structural similarity between the target protein monomers and the protein-protein complexes previously determined by experimental techniques (templates). The template-based docking relies on the quality and diversity of the template set. We present a carefully curated, nonredundant library of templates containing 4950 full structures of binary complexes and 5936 protein-protein interfaces extracted from the full structures at 12 Å distance cut-off. Redundancy in the libraries was removed by clustering the PDB structures based on structural similarity. The value of the clustering threshold was determined from the analysis of the clusters and the docking performance on a benchmark set. High structural quality of the interfaces in the template and validation sets was achieved by automated procedures and manual curation. The library is included in the Dockground resource for molecular recognition studies at http://dockground.bioinformatics.ku.edu.

Keywords: benchmark sets; protein interactions; protein modeling; protein recognition; structure prediction.

Figure 1

Flowchart of algorithm for generation of full-structure and interface template libraries.

Figure 2

Example of a “bad” complex. Chains D and E from 1fma are shown in magenta and yellow, respectively, with penetrating chain removed by the automated procedure described in the text. Buried Val and Thr residues at the C-terminal of chain F identified by the procedure are shown as sticks (the last two residues at the terminus are Gly).

Figure 3

Properties of similarity graphs. (A) Protein-protein complexes and (B) protein-protein interfaces. The main panels show distributions of connected component size at different thresholds of the clustering TM-score (TM_T). The inserts display dependence of clustering coefficient CC on TM_T.

Figure 4

Correlation of protein-protein and interface-interface TM-scores.

Figure 5

Number of connected components and clusters as a function of clustering threshold. (A) Protein-protein complexes, and (B) protein-protein interfaces. N_CC is the number of connected components, and N_C is the number of clusters. Green lines (scaled to the right-hand axes) show the relative increase in the number of connected graph parts after splitting the connected components into tightly connected clusters.

Figure 6

Quality of clusters at different clustering thresholds. (A) Protein-protein complexes, and (B) protein-protein interfaces. Distributions of missing edges per cluster are shown as box-and-whickers plots with horizontal lines for minimal, maximal and median values in the distributions and boxes containing second and third quartiles of data. The circles (scaled to the right hand axis) show how the fraction of clusters, which are not complete sub-graphs of the initial similarity graphs, depends on TM_T.

Figure 7

Performance of structure alignment at different clustering thresholds. Full-structure (circles) and interface (triangles) libraries were generated at different threshold values. Success rates were calculated for the entire pool of structures excluding templates similar to the target (see text).