AI and Machine Learning for Proteomics-Driven Drug Discovery: Methods, Tools, and Best Practices
- 1Department of Chemistry, Technical University of Denmark, 2800 Kongens Lyngby, Denmark
- 2Department of Health Technology (DTU Health Technology), Technical University of Denmark, 2800 Kongens Lyngby, Denmark
Abstract
Proteomics has become central to pharmacological research by providing quantitative readouts of protein abundance, post-translational modifications, interactions, and spatial context. However, proteomic datasets are high-dimensional, heterogeneous, and frequently affected by missingness, batch effects, and limited cohort size. Artificial intelligence (AI) and machine learning (ML) can help convert these complex data into decision-relevant outputs for target identification, biomarker discovery, pharmacodynamic monitoring, and drug repurposing. This review critically compares supervised learning, ensemble methods, dimensionality reduction, clustering, deep learning, graph learning, survival modeling, causal inference, and calibration approaches in proteomics-driven drug discovery. We also summarize major software ecosystems for mass-spectrometry processing, targeted assays, spectrum prediction, phosphoproteomics, structure modeling, and reproducible workflows. Emphasis is placed on model selection, benchmarking, missing-data handling, batch correction, interpretability, uncertainty, experimental validation, and translational readiness. Finally, we highlight emerging directions, including contrastive learning, diffusion models, graph-based integration, and federated analytics.
Related
- The NERDD microservices platform for deploying machine learning models in drug discovery
- Machine-learning-guided discovery of selective CK2 inhibitors with reduced CDK off-target activity
- Integrated proteomics and machine learning for identifying candidate serum biomarkers in acute myocardial infarction-complicated cardiogenic shock: a prospective exploratory study
- MLIP Detective: Active Failure Mode Discovery Beyond Benchmark Scores for Machine-Learning Interatomic Potentials
- Discovery of germanium-based hybrid perovskites with high photovoltaic performance with machine-learning
BibTeX
@article{basak2026machine,
title = {{AI and Machine Learning for Proteomics-Driven Drug Discovery: Methods, Tools, and Best Practices}},
author = {Suman Basak},
journal = {Current Issues in Molecular Biology},
year = {2026},
month = may,
doi = {10.3390/cimb48050532},
url = {https://doi.org/10.3390/cimb48050532},
note = {CC BY},
}
RIS
TY - JOUR
TI - AI and Machine Learning for Proteomics-Driven Drug Discovery: Methods, Tools, and Best Practices
AU - Suman Basak
PY - 2026
DA - 2026/05/20
JO - Current Issues in Molecular Biology
DO - 10.3390/cimb48050532
UR - https://doi.org/10.3390/cimb48050532
AB - Proteomics has become central to pharmacological research by providing quantitative readouts of protein abundance, post-translational modifications, interactions, and spatial context. However, proteomic datasets are high-dimensional, heterogeneous, and frequently affected by missingness, batch effects, and limited cohort size. Artificial intelligence (AI) and machine learning (ML) can help convert these complex data into decision-relevant outputs for target identification, biomarker discovery, pharmacodynamic monitoring, and drug repurposing. This review critically compares supervised learning, ensemble methods, dimensionality reduction, clustering, deep learning, graph learning, survival modeling, causal inference, and calibration approaches in proteomics-driven drug discovery. We also summarize major software ecosystems for mass-spectrometry processing, targeted assays, spectrum prediction, phosphoproteomics, structure modeling, and reproducible workflows. Emphasis is placed on model selection, benchmarking, missing-data handling, batch correction, interpretability, uncertainty, experimental validation, and translational readiness. Finally, we highlight emerging directions, including contrastive learning, diffusion models, graph-based integration, and federated analytics.
SN - 1467-3037
SN - 1467-3045
ER -