Assay-Aware BindingDB: Curating Experimental Context for Binding Affinity Prediction

Abstract

Protein--ligand binding affinity prediction is fundamental to computational drug discovery, yet modern AI-driven models are limited by pervasive heterogeneity in their training data: bioactivity values are aggregated across diverse assay types and experimental conditions without accounting for protocol-level differences, introducing systematic noise. Existing harmonization approaches either discard assay-level metadata or collapse it into coarse categorical distinctions, leaving rich contextual signal unused. We address this gap with two contributions. First, we introduce Assay-Aware BindingDB, augmenting 74,425 BindingDB protein--ligand pairs across four assay types (ITC, SPR, RBA, and FPA) with structured metadata extracted from primary literature using a two-stage agentic framework that separates evidence extraction from ontology-conditioned JSON synthesis. Against domain-expert references, the framework achieves presence F1 ≥ 0.947 and semantic content accuracy ≥ 0.913 across all assay types. Second, we develop a context-conditioned affinity model that injects an assay-context embedding into the Boltz-2 affinity module. On a paper-level held-out split, the model reduces combined-regime MSE from $1.27$ to $1.19$ and increases Pearson correlation from $0.64$ to $0.67$, with significant gains for SPR and FPA. ITC, the only label- and immobilization-free assay considered, shows no improvement, consistent with its design removing the protocol artifacts the metadata captures. These results support the hypothesis that systematically curated assay metadata provides informative signal for affinity prediction.

Systematic comparison of affinity plasma proteomic technologies across multiple cohorts reveals generalizability of proteomic prediction
Anthony Barente, Zijie Zhao, Zhili Zheng +16
Nature Communications8 Oct 2026Peer-reviewedCC BY-NC-ND 4.0Full textMolecular biology & gene regulationCrossref
Chemical and geometric representation fidelity improves drug--target affinity prediction
Yixiao Li, Yining Qian, Yefan Chen +5
arXiv2 Sep 2026PreprintarXiv licenceFull textBiophysics StatisticsarXiv
CP-GFM-SGX: Topology-Aware Cyclic Peptide Generation and Design with Receptor-Conditioned Structural Evaluation and Experimental KRAS Affinity Optimization
Zhiyuan Hu, Russell Dillworth
ChemRxiv12 Oct 2026PreprintCC BYFull textMolecular biology & gene regulationChemRxiv
PHL: Persistent Hyperdigraph Learning for Protein-Protein Binding Affinity Prediction
Xingjian Xu, Chunmei Wang, Jiahui Chen
arXiv27 Sep 2026PreprintCC BY 4.0Full textBiophysicsarXiv
PSLL: Persistent Sheaf Laplacian Learning for Protein-Ligand Binding Affinity Prediction
Mushal Zia, Benjamin Jones, Guo-Wei Wei
arXiv1 Sep 2026PreprintCC BY 4.0Full textBiophysicsarXiv

Cite

BibTeX RIS

BibTeX

@misc{wu2026assayaware,
  title = {{Assay-Aware BindingDB: Curating Experimental Context for Binding Affinity Prediction}},
  author = {Ming-Hsiu Wu and Xuejiao Shirley Guo and Bingsong Zeng and Ziqian Xie and Shuiwang Ji and Wenshe Liu and Cui Tao and Degui Zhi},
  howpublished = {arXiv},
  year = {2026},
  month = sep,
  doi = {10.48550/arxiv.2609.34001},
  eprint = {2609.34001},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2609.34001},
  note = {CC BY 4.0},
}

RIS

TY  - UNPB
TI  - Assay-Aware BindingDB: Curating Experimental Context for Binding Affinity Prediction
AU  - Ming-Hsiu Wu
AU  - Xuejiao Shirley Guo
AU  - Bingsong Zeng
AU  - Ziqian Xie
AU  - Shuiwang Ji
AU  - Wenshe Liu
AU  - Cui Tao
AU  - Degui Zhi
PY  - 2026
DA  - 2026/09/27
PB  - arXiv
DO  - 10.48550/arxiv.2609.34001
UR  - https://arxiv.org/abs/2609.34001
AB  - Protein--ligand binding affinity prediction is fundamental to computational drug discovery, yet modern AI-driven models are limited by pervasive heterogeneity in their training data: bioactivity values are aggregated across diverse assay types and experimental conditions without accounting for protocol-level differences, introducing systematic noise. Existing harmonization approaches either discard assay-level metadata or collapse it into coarse categorical distinctions, leaving rich contextual signal unused. We address this gap with two contributions. First, we introduce Assay-Aware BindingDB, augmenting 74,425 BindingDB protein--ligand pairs across four assay types (ITC, SPR, RBA, and FPA) with structured metadata extracted from primary literature using a two-stage agentic framework that separates evidence extraction from ontology-conditioned JSON synthesis. Against domain-expert references, the framework achieves presence F1 ≥ 0.947 and semantic content accuracy ≥ 0.913 across all assay types. Second, we develop a context-conditioned affinity model that injects an assay-context embedding into the Boltz-2 affinity module. On a paper-level held-out split, the model reduces combined-regime MSE from $1.27$ to $1.19$ and increases Pearson correlation from $0.64$ to $0.67$, with significant gains for SPR and FPA. ITC, the only label- and immobilization-free assay considered, shows no improvement, consistent with its design removing the protocol artifacts the metadata captures. These results support the hypothesis that systematically curated assay metadata provides informative signal for affinity prediction.
ER  -

Abstract

Related

Cite