Comparing protein structures with RINspector automation in Cytoscape

Abstract

Residue interaction networks (RINs) have been shown to be relevant representations of the tertiary or quaternary structures of proteins, in particular thanks to network centrality analyses. We recently developed the RINspector Cytoscape app, which couples centrality analyses with backbone flexibility predictions. This combined approach permits the identification of crucial residues for the folding or function of the protein that can constitute good targets for mutagenesis experiments. Here we present an application programming interface (API) for RINspector that enables interplay between Cytoscape, RINspector and external languages, such as R or Python. This API provides easy access to batch centrality calculations and flexibility predictions, and allows for the easy comparison of results between different structures. These comparisons can lead to the identification of specific and conserved central residues, and show the impact of mutations to these and other residues on the flexibility of the proteins. We give two use cases to demonstrate the interest of these functionalities and provide the corresponding scripts: the first concerns NMR conformers, the second focuses on mutations in a structure.

The Inclusion of Water Molecules in Residue Interaction Networks Identifies Additional Central Residues

“…We computed Residue Centrality Analyses (RCAs) and Betweenness Centrality Analyses (BCAs) on all the RINs with the RINspector app (Brysbaert et al, 2018a ) for Cytoscape using its automation functionalities (Brysbaert et al, 2018b ). Each analysis produces a Z-score for each residue.…”

“…The question we address in this paper is how the inclusion of protein-associated water molecules affects the functional importance of protein residues. By employing a tool we recently developed (Brysbaert et al, 2018a , b ), we consider the residue interaction networks of exemplary protein complexes under inclusion of surface and interfacial water molecules. Specifically, we determine what effect the water molecules, when considered as nodes in these networks, have on centrality calculations of protein residues, i.e., whether and in what way they modify the results obtained for the waterless proteins, in order to gain additional insight into the biological role they play for the relevance of protein residues.…”

Identification of Key Residues in Proteins Through Centrality Analysis and Flexibility Prediction with RINspector

“…It becomes a bit laborious when several structures must be analyzed sequentially, as in the case of NMR data, for instance. Since its release 3.3, Cytoscape embeds a tool named CyREST (Ono, Muetze, Kolishovski, Shannon, & Demchak, 2015) that enables the user to call functions of Cytoscape from an external program written in, e.g., Python or R. RINspector in its release 1.1.0 permits such a call of its centrality analyses and DynaMine prediction functions (Brysbaert, Mauri, & Lensink, 2018). This protocol describes how to call RINspector functions from a script in Python or R to perform calculations on the four chains of the PDB 4R6E.…”

“…The flexibility prediction of each chain is thus obtained. More information about automation and advanced scripts for RINspector is available in Brysbaert et al (2018).…”

“…These structures can be used for modeling at different levels, like homology modeling, protein‐ligand docking, protein‐protein docking, dynamic predictions, and simulations, or used to identify regions or residues of interest that may lead to mutagenesis experiments or be employed for targeted drug design. We have recently developed a tool called RINspector (Brysbaert, Lorgouilloux, Vranken, & Lensink, , ), the goal of which is to identify key residues in a structure that are amenable to mutagenesis experiments. It is included as an app in the Cytoscape network visualization and analysis software (Shannon et al., ; Su, Morris, Demchak, & Bader, ), accessible through its apps manager.…”