Skip to content
DraconiticResearch
Search Sources API
197,768 studies31 sourcesUpdated 21 min ago
PreprintCC BYChemRxiv

Guidelines to Evaluate Structure-based Methods for Protein-Ligand Binding Prediction

  • Viet-Khoa Tran-Nguyen1,2
  • Pedro J. Ballester3,4
4 affiliations
  1. 1Université Paris Cité
  2. 2INSERM ERL
  3. 3Department of Bioengineering
  4. 4Imperial College London
Open at sourcePDFDOICite

Abstract

Scoring functions built with advanced deep learning algorithms have recently gained prominence in protein-ligand pose generation and binding affinity prediction for computational drug design. However, a number of shortcomings make it increasingly difficult to evaluate scoring functions for a specific task across studies. Transparent disclosure of training data, code and procedures is essential to permit reproducibility, reuse and independent scrutiny. Guarding against leakage, sampling bias and flawed design is equally important. Furthermore, meaningful comparisons require common data splits, checks for run-to-run variability and performance baselines. Here we propose a set of guidelines to help the community identify the most promising methods for these important problems.

Related

  1. Reliability of AI Methods in Drug Discovery: Evaluation of Boltz-2 for Structure and Binding Affinity Prediction
    Shunzhou Wan, Xibei Zhang, Xue Xiao +1
    Journal of Chemical Theory and ComputationPeer-reviewedCC BYFull text20 citedMedicinal chemistry & drug discoveryOpenAlex
  2. Logarithmic-scale variational quantum eigensolver for off-lattice protein structure prediction in continuous torsional angle space
    Fabio Cumbo, Bryan Raubenolt, Varun Puram +3
    arXivPreprintCC BY 4.0Full textPhysical chemistryQuantum physics & quantum informationarXiv
  3. Using structure to evaluate the accuracy and transferability of machine learning potentials of biomolecules
    Lejla Š. Biberić, Nisarg Joshi, Jim Pfaendtner
    RSC AdvancesPeer-reviewedCC BY-NC 3.0Full textMaterials & polymersPMC
  4. Synthetic Mechanism of a Fe(II) N -Heterocyclic Carbene Bidentate Complex Revealed by Electronic Structure Methods
    Abdelazim Abdelgawwad, Ulises Carrillo, Philippe C. Gros +2
    Inorganic ChemistryPeer-reviewedCC BY 4.0Full textOrganic chemistryHAL
  5. 3DRISM Solvation for Accurate and Large-Scale Protein Hydration Prediction and Mechanistic Interpretation
    Peter C. Swanson, Siqin Cao, Zige Liu +3
    ChemRxivPreprintLicence not statedGeneral chemistryChemRxiv

Cite

BibTeXRIS
BibTeX
@misc{trannguyen2026guidelines,
  title = {{Guidelines to Evaluate Structure-based Methods for Protein-Ligand Binding Prediction}},
  author = {Viet-Khoa Tran-Nguyen and Pedro J. Ballester},
  howpublished = {ChemRxiv},
  year = {2026},
  month = sep,
  doi = {10.26434/chemrxiv.15008174/v1},
  url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008174/v1},
  note = {CC BY},
}
RIS
TY  - UNPB
TI  - Guidelines to Evaluate Structure-based Methods for Protein-Ligand Binding Prediction
AU  - Viet-Khoa Tran-Nguyen
AU  - Pedro J. Ballester
PY  - 2026
DA  - 2026/09/01
PB  - ChemRxiv
DO  - 10.26434/chemrxiv.15008174/v1
UR  - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008174/v1
AB  - Scoring functions built with advanced deep learning algorithms have recently gained prominence in protein-ligand pose generation and binding affinity prediction for computational drug design. However, a number of shortcomings make it increasingly difficult to evaluate scoring functions for a specific task across studies. Transparent disclosure of training data, code and procedures is essential to permit reproducibility, reuse and independent scrutiny. Guarding against leakage, sampling bias and flawed design is equally important. Furthermore, meaningful comparisons require common data splits, checks for run-to-run variability and performance baselines. Here we propose a set of guidelines to help the community identify the most promising methods for these important problems.
ER  -