DeepKinomeWeb: a quantitative, panel-level platform for kinase inhibitor screening and selectivity profiling
Clinical Snapshot
PICO Framework
| P — Population | Protein kinases as molecular targets in preclinical drug discovery; intended end-users are biologists and medicinal chemists engaged in kinase inhibitor development |
| I — Intervention | DeepKinomeWeb — a freely accessible, web-based platform integrating a deep learning regression model (DeepKinome) for quantitative prediction of kinase-inhibitor binding affinities, panel-level selectivity visualisation, selectivity metric calculations, and structural/physicochemical analyses |
| C — Comparator | No formal head-to-head comparator is explicitly described; implicit comparison is against conventional high-throughput experimental screening and existing computational tools for kinase selectivity profiling |
| O — Outcomes | Predictive accuracy of kinase-inhibitor binding affinity (quantitative), selectivity landscape characterisation at panel level, interpretability of large-scale screening data, and usability for rational drug discovery decision-making |
Bottom Line
DeepKinomeWeb is a freely accessible, web-based computational platform designed to predict kinase-inhibitor binding affinities and profile selectivity across the human kinome, built upon the previously validated DeepKinome deep learning model. It addresses a genuine bottleneck in early-stage drug discovery by transforming large-scale competition-based screening data into interpretable, actionable insights for medicinal chemists and biologists. The platform integrates quantitative binding affinity prediction, selectivity metric calculations, and structural and physicochemical analyses in a single interface — a meaningful advance in workflow integration. However, the published abstract does not report quantitative validation metrics, confidence intervals, or head-to-head benchmarking against existing tools, making independent assessment of predictive accuracy impossible at this stage. Key methodological questions — including kinome coverage of training data, handling of assay heterogeneity, and performance on novel chemical scaffolds — remain unanswered. For Australian researchers, the tool is freely accessible and potentially valuable for prioritising kinase inhibitor candidates before costly experimental screening, but all computational predictions must be experimentally validated before progression to in vivo or clinical development. Clinicians should regard this as a promising preclinical research tool rather than a validated clinical decision-support instrument.
Key Findings
P Value: Not reported
Effect Size: Not reported in abstract; referenced to prior DeepKinome validation publication
Primary Outcome: Quantitative prediction of kinase-inhibitor binding affinities and panel-level selectivity landscape visualisation via a web-based interface; specific predictive performance metrics are not reported in the abstract
Nnt Or Sensitivity: Not applicable (preclinical computational tool); no sensitivity, specificity, or NNT data reported. Relevant metrics for this study type would include RMSE, Pearson/Spearman correlation coefficients, and selectivity score concordance — none are provided in the abstract
Confidence Interval: Not reported
Clinical Application
High feasibility for research adoption: the platform is freely accessible, requires no login, and is designed with a user-friendly interface. Integration into existing drug discovery workflows is straightforward for teams with basic computational literacy. Experimental validation of computational predictions remains essential before advancing candidates to in vivo or clinical stages. DeepKinomeWeb has indirect relevance to Australian healthcare through its potential to accelerate preclinical kinase inhibitor discovery. Australia has an active kinase inhibitor research landscape, with TGA-approved kinase inhibitors (e.g., imatinib, ibrutinib, osimertinib) listed on the PBS for haematological malignancies and solid tumours. Australian academic drug discovery programs (e.g., through WEHI, QIMR Berghofer, and university medicinal chemistry groups) could utilise this freely available tool to prioritise inhibitor candidates before costly experimental screening. However, TGA regulatory submissions require robust experimental validation data; computational predictions from DeepKinomeWeb would serve as hypothesis-generating tools rather than regulatory evidence. RACGP guidelines do not directly address preclinical computational tools, but the platform's outputs are relevant to oncology and precision medicine pipelines that ultimately inform PBS listing decisions. Preclinical researchers, medicinal chemists, and computational biologists engaged in kinase inhibitor drug discovery programs; not directly applicable to clinical patient populations at this stage
Abstract
Protein kinases are central targets in drug discovery, yet early-stage development of potent and selective inhibitors remains challenging due to high experimental costs and limited interpretability of large-scale screening data. Here, we present DeepKinomeWeb, an integrated web-based platform that transforms competition-based high-throughput screening data into actionable insights for kinase inhibitor prioritization. Built upon our previously validated deep learning regression model, DeepKinome, the platform enables quantitative prediction of kinase-inhibitor binding affinities and provides panel-level visualization of selectivity landscapes, selectivity metric calculations, and integrated structural and physicochemical analyses. Through its user-friendly interface, DeepKinomeWeb supports rational, data-driven decision-making for biologists and medicinal chemists, lowering the barrier to systematic selectivity assessment in kinase inhibitor discovery. DeepKinomeWeb is freely available to all users without any login requirement at https://str.kribb.re.kr/deepkinome.
References
- 1.Eun, J., Lee, Y., Yang, S., Choi, D., Na, H., Cho, H., Nam, S., & Lee, J. (2026). DeepKinomeWeb: a quantitative, panel-level platform for kinase inhibitor screening and selectivity profiling. Nucleic Acids Research. https://doi.org/10.15252/embr.201643300
Related Research
The American journal of pathology
An Integrated Machine Learning and Genomic Framework for Precise Detection of Gastric Cancer
3 Aug 2026
Current opinion in chemical biology
Prediction of protein-protein interactions and co-complex models with deep learning
3 Aug 2026
European journal of radiology
A CT-based deep learning model to differentiate between benign and malignant adrenal lesions
3 Aug 2026
This content is for educational purposes for healthcare professionals only and does not constitute clinical advice. Clinical decisions should be based on individual patient assessment, current guidelines, and appropriate specialist consultation. Editorial Standards · Privacy Policy · Terms of Service