Oct 2026· Journal of Chemical Information and Modeling· Vol 66 20, pp.
13618-13629
· 0 citations· 31 references
Medicine
TL;DR
The results establish that diversity-aware search broadens the set of docking-predicted interaction hypotheses generated under a fixed budget; they do not experimentally establish distinct binding modes or biochemical activity.
Abstract
Structure-based discovery pipelines commonly prioritize a scalar docking score, concentrating candidates into narrow chemotypes and contact patterns. We formulate docking-guided de novo molecular design as a diversity-seeking exploration problem and compare goal-directed, quality-diversity, Pareto-optimization, affinity-focused, novelty-based, and random search policies. Molecules are generated by fragment-based CReM edits and scored by rigid-receptor GNINA/Vina docking. We introduce a residue-level multi-interaction fingerprint encoding hydrophobic contacts and hydrogen-bond donor and acceptor interactions, together with the reachability ratio, an empirical diagnostic of goal-retrieval discriminability. Across seven targets from five protein families, with 500 docking evaluations per run, Curiosity-IMGEP discovers 15.0% more interaction profiles than random search (d = 1.01, pBonf < 10-6) and produces +12% more Bemis-Murcko scaffolds and +21% more binary residue-contact patterns. A newly added budget-matched NSGA-II baseline achieves the strongest predicted binding affinity in the benchmark (-11.03 vs Genetic algorithm's -10.98 kcal/mol; d = -1.02 vs Random, pBonf < 10-8), tied within noise with the Genetic algorithm, and leads on Bemis-Murcko scaffold count (23 ± 12), but underperforms QD/goal-directed methods on interaction-profile counts (259 ± 46, -12% vs Random). Representation ablations compare continuous and binary fingerprints, charged-interaction channels, distance cutoffs, and inclusion of docking quality in the search vector. Target-specific ChEMBL neighborhood and pose-robustness analyses provide orthogonal computational triage. The results establish that diversity-aware search broadens the set of docking-predicted interaction hypotheses generated under a fixed budget; they do not experimentally establish distinct binding modes or biochemical activity. Code, configurations, and raw results are publicly available.
This work developed a compound prioritization strategy that utilizes sampled docked poses from complementary docking approaches into multiple conformations of the target protein to train two multilayer perceptron (MLP) models, and found that a combination of scoring and prediction metrics enriches for hit compounds acr...
A. Chang-Gonzalez, Alexis N. Campbell, Eric W. Bell et al.· Journal of Chemical Informat...· 0 citations
Dipeptidyl peptidase-4 (DPP4) is a validated target for type 2 diabetes, and virtual screening of natural product libraries is a common route to new chemotypes. We asked whether a top docking rank survives independent scrutiny. From COCONUT, 85,600 compounds were filtered to 32,008 and docked against DPP4; 873 outscore...
E. Aktaş· SAR and QSAR in environmenta...· 0 citations
Introduction Host-directed antiviral discovery requires a traceable route from host-factor identification to testable molecular hypotheses. This study developed such a workflow around a host factor associated with Ebola virus glycoprotein (EBOV GP). Methods GFP-GP pulldown coupled with co-immunoprecipitation/mass spect...
Yan-Shuang Wang, Yuan-Jian Li, Lin Zhu et al.· Frontiers in Pharmacology· 0 citations
The results support a staged screening design in which a pooled default is followed by ligand-specific calibration and expert review, and SCScore showed the highest association with the blinded consensus, but performance varied by ligand.
T. Sobodu, Victoria Johnson, Ryan Kern et al.· 0 citations
Predicting how a ligand binds to its target is central to small-molecule drug discovery. Modern cofolding models can generate diverse candidate poses, but reliably ranking these candidates remains challenging. We introduce Pyrite, a pairwise pose-ranking benchmark comprising 1,500 pose pairs across diverse protein-liga...
Megha Srivastava, Michael H. Lim, Allison M. Keys et al.· 1 citation