PrankWeb 3: accelerated ligand-binding site predictions for experimental and modelled protein structures

• Published in: Nucleic acids research Vol. 50; no. W1; pp. W593 - W597
• Main Authors: Jakubec, David, Skoda, Petr, Krivak, Radoslav, Novotny, Marian, Hoksza, David
• Format: Journal Article
• Language: English
• Published: England Oxford University Press 05-07-2022
• Subjects: Binding Sites; Databases, Protein; Internet; Ligands; Protein Binding; Protein Domains; Proteins - chemistry; Software; Web Server Issue
• ISSN: 0305-1048; 1362-4962
• DOI: 10.1093/nar/gkac389

Abstract
Knowledge of protein–ligand binding sites (LBSs) enables research ranging from protein function annotation to structure-based drug design. To this end, we have previously developed a stand-alone tool, P2Rank, and the web server PrankWeb (https://prankweb.cz/) for fast and accurate LBS prediction. Here, we present significant enhancements to PrankWeb. First, a new, more accurate evolutionary conservation estimation pipeline based on the UniRef50 sequence database and the HMMER3 package is introduced. Second, PrankWeb now allows users to enter UniProt ID to carry out LBS predictions in situations where no experimental structure is available by utilizing the AlphaFold model database. Additionally, a range of minor improvements has been implemented. These include the ability to deploy PrankWeb and P2Rank as Docker containers, support for the mmCIF file format, improved public REST API access, or the ability to batch download the LBS predictions for the whole PDB archive and parts of the AlphaFold database.
Graphical Abstract
Graphical Abstract
PrankWeb accepts a protein structure on its input, computes evolutionary conservation, and predicts binding sites which are then mapped onto the structure and can be visually inspected.