Skip to content
DraconiticResearch
Search Sources API
197,768 studies31 sourcesUpdated 10 min ago
Peer-reviewedCC BY 4.0Mathematics

Metaheuristic Optimization of Temporal Surrogates for Early Prediction of Protein Conformational Stability from Short Molecular-Dynamics Trajectories

  • Francisco Hernando-Gallego1
  • Diego García-Santos2
  • Dimitar Vasilev-Vasilev2
  • Diego Martín2
  • José Vicente Álvarez-Bravo2
  1. 1Department of Applied Mathematics, Escuela de Ingeniería Informática de Segovia, Universidad de Valladolid, 40005 Segovia, Spain
  2. 2Department of Computer Science, Escuela de Ingeniería Informática de Segovia, Universidad de Valladolid, 40005 Segovia, Spain
Open at sourceDOICite

Abstract

Metaheuristic optimization can strengthen recurrent surrogates, but its value must be judged against controls and hyperparameter optimizers under leakage-free evaluation. We predict five equilibrium protein observables from 20 ns prefixes of 100 ns molecular-dynamics trajectories using a gated recurrent unit (GRU) configured by an Improved Grey Wolf Optimizer (Improved-GWO). The benchmark comprises 20 ATLAS proteins (60 trajectories) under nested leave-one-protein-out cross-validation. Relative to the original fixed GRU, Improved-GWO had lower absolute error for several observables; Holm-adjusted paired tests remained below 0.05 for SASA, helix and strand. Among six models, last-value persistence achieved the lowest nominal MAE for all five endpoints, and no persistence–Improved-GWO comparison survived Holm correction. A fixed 512-unit GRU improved nominally over the original small GRU, while paired tests favored Improved-GWO for helix and strand. In matched-budget HPO, Bayesian Optimization attained lower best inner-validation fitness than Improved-GWO, which in turn was lower than Random Search. Configurations from Random Search and Bayesian Optimization were not evaluated on outer tests. Ridge remained competitive. Thus, the benchmark does not establish a consistent predictive advantage from recurrent complexity over temporal or regularized linear baselines. The p53/NAMD assessment remains preliminary zero-shot evidence, and trajectory-length reduction is not a measured wall-clock speedup or validated stopping policy.

Related

  1. MathAgent: Multi-Agent Optimization of Mathematical Invariants for Molecular Property Prediction
    Yiming Ren, Josh V. Vermaas, Guo-Wei Wei
    arXivPreprintarXiv licenceFull textBiophysicsarXiv
  2. RFOptimization: Guiding Design Optimization with All-Atom Structure Prediction
    O. Zhang, J. Wang, T. R. Thompson +4
    bioRxivPreprintCC BYFull textMolecular biology & gene regulationbioRxiv
  3. A Leakage-Controlled Benchmark of Metaheuristic Multi-Omics Feature Selection for Reconstructing TCGA Molecular Subtypes in Endometrial Cancer
    Minyue Cao, Yiqin Zhang, Yan Ding +7
    CancersPeer-reviewedCC BY 4.0Molecular biology & gene regulationCrossref
  4. Flat Amino Acid Surrogates Induce Turn Conformers: Synthesis, Nitrogen Chirality, Conformation and Biomedical Applications of N-Amino Triazolone (Nat) Peptides
    Chitra P. Sadanandhan, Ida Boccino, Mukandila Mulumba +8
    ChemRxivPreprintCC BY-NC-NDFull textMolecular biology & gene regulationChemRxiv
  5. Exploring the conformational space of membrane protein folds matching distance constraints.
    Jean-Loup Faulon, Ken Sale, Malin Young
    Protein SciencePeer-reviewedLicence not stated24 citedMolecular biology & gene regulationHAL

Cite

BibTeXRIS
BibTeX
@article{hernandogallego2026metaheuristic,
  title = {{Metaheuristic Optimization of Temporal Surrogates for Early Prediction of Protein Conformational Stability from Short Molecular-Dynamics Trajectories}},
  author = {Francisco Hernando-Gallego and Diego García-Santos and Dimitar Vasilev-Vasilev and Diego Martín and José Vicente Álvarez-Bravo},
  journal = {Mathematics},
  year = {2026},
  month = oct,
  doi = {10.3390/math14203679},
  url = {https://www.mdpi.com/2227-7390/14/20/3679},
  note = {CC BY 4.0},
}
RIS
TY  - JOUR
TI  - Metaheuristic Optimization of Temporal Surrogates for Early Prediction of Protein Conformational Stability from Short Molecular-Dynamics Trajectories
AU  - Francisco Hernando-Gallego
AU  - Diego García-Santos
AU  - Dimitar Vasilev-Vasilev
AU  - Diego Martín
AU  - José Vicente Álvarez-Bravo
PY  - 2026
DA  - 2026/10/10
JO  - Mathematics
DO  - 10.3390/math14203679
UR  - https://www.mdpi.com/2227-7390/14/20/3679
AB  - Metaheuristic optimization can strengthen recurrent surrogates, but its value must be judged against controls and hyperparameter optimizers under leakage-free evaluation. We predict five equilibrium protein observables from 20 ns prefixes of 100 ns molecular-dynamics trajectories using a gated recurrent unit (GRU) configured by an Improved Grey Wolf Optimizer (Improved-GWO). The benchmark comprises 20 ATLAS proteins (60 trajectories) under nested leave-one-protein-out cross-validation. Relative to the original fixed GRU, Improved-GWO had lower absolute error for several observables; Holm-adjusted paired tests remained below 0.05 for SASA, helix and strand. Among six models, last-value persistence achieved the lowest nominal MAE for all five endpoints, and no persistence–Improved-GWO comparison survived Holm correction. A fixed 512-unit GRU improved nominally over the original small GRU, while paired tests favored Improved-GWO for helix and strand. In matched-budget HPO, Bayesian Optimization attained lower best inner-validation fitness than Improved-GWO, which in turn was lower than Random Search. Configurations from Random Search and Bayesian Optimization were not evaluated on outer tests. Ridge remained competitive. Thus, the benchmark does not establish a consistent predictive advantage from recurrent complexity over temporal or regularized linear baselines. The p53/NAMD assessment remains preliminary zero-shot evidence, and trajectory-length reduction is not a measured wall-clock speedup or validated stopping policy.
SN  - 2227-7390
ER  -