pocketz.doi.bio/2026-05-13_131409_researchgatenet_374489672_Assessment_of_protein-ligand_complexes_in_CASP15

Assessment of protein–ligand complexes in CASP15 | Request PDF

URL: https://www.researchgate.net/publication/374489672Assessmentofprotein-ligandcomplexesinCASP15

Saved: 2026-05-13T20:14:14.377Z

Page 1 R E S E A R C H A R T I C L E Assessment of protein – ligand complexes in CASP15 Xavier Robin 1,2

Gabriel Studer 1,2

Janani Durairaj 1,2

Jerome Eberhardt 1,2

| Torsten Schwede 1,2

W. Patrick Walters 3 1Biozentrum, University of Basel, Basel, Switzerland 2SIB Swiss Institute of Bioinformatics, Basel, Switzerland 3Relay Therapeutics, Cambridge, Massachusetts, USA Correspondence W. Patrick Walters, Relay Therapeutics, Cambridge, MA, USA. Email: pwalters@relaytx.com Abstract CASP15 introduced a new category, ligand prediction, where participants were pro- vided with a protein or nucleic acid sequence, SMILES line notation, and stoichiome- try for ligands and tasked with generating computational models for the three- dimensional structure of the corresponding protein – ligand complex. These models were subsequently compared with experimental structures determined by x-ray crys- tallography or cryoEM. To assess these predictions, two novel scores were devel- oped. The Binding-Site Superposed, Symmetry-Corrected Pose Root Mean Square Deviation (BiSyRMSD) evaluated the absolute deviations of the models from the experimental structures. At the same time, the Local Distance Difference Test for Protein – Ligand Interactions (lDDT-PLI) assessed the ability of models to reproduce the protein – ligand interactions in the experimental structures. The ligands evaluated in this challenge range from single-atom ions to large flexible organic molecules. More than 1800 submissions were evaluated for their ability to predict 23 different protein – ligand complexes. Overall, the best models could faithfully reproduce the geometries of more than half of the prediction targets. The ligands' size and flexibility were the primary factors influencing the predictions' quality. Small ions and organic molecules with limited flexibility were predicted with high fidelity, while reproducing the binding poses of larger, flexible ligands proved more challenging. K E Y W O R D S CASP, protein structure prediction, protein – ligand complex 1

I N T RO DU CT I O N CASP15 included a new category that assessed the ability of compu- tational methods to predict not only the structures of proteins but also the orientations and conformations of bound ligands (commonly referred to as binding poses). Like previous CASP protein structure prediction challenges, participants were provided with the sequence for the protein or RNA structures. The procedure, submissions, and formats are described in more detail in a separate paper in this issue.1 In addition, the structures and stoichiometries of bound ligands were made available in the Simplified Molecular Input Line Entry System (SMILES)2,3 notation. The

“ significance ”

of the ligand in the structure was also included as a field in the input. Only ligands where the significance was labeled as

“ Yes ”

in the input were assessed. Given a sequence, stoichiometry, and SMILES, participants were tasked with modeling the structure of the protein – ligand complex. The ligand prediction challenge in CASP15 differed from past ligand prediction challenges like Teach Discover Treat (TDT),4 Continu- ous Evaluation of Ligand Prediction Performance (CELPP),5 Drug Dis- covery Data Resource (D3R),6 – 9 and Community Structure – Activity Resource (CSAR).10,11 In previous challenges, an experimental protein structure was provided, and the participants were simply required to predict a docked ligand pose. In addition, these challenges provided information regarding the ligand's binding site. In CASP15, participants were responsible for generating the protein structure and identifying the ligand binding sites as well as the binding poses of the ligands. Received: 11 April 2023

Revised: 14 September 2023

Accepted: 19 September 2023 DOI: 10.1002/prot.26601 Proteins.

2023;91:1811 – 1821.

wileyonlinelibrary.com/journal/prot

1811