Work

Hybrid-fingerprint machine learning framework for predicting inhibitors relevant to breast cancer

Year: 2026

Type: article

Abstract: Breast cancer remains a leading cause of cancer morbidity and mortality, motivating rapid discovery of targeted therapeutics that modulate hormone biosynthesis and metastatic progression. Among clinically relevant molecular targets are aromatase (CYP19A1) and matrix metalloproteinases MMP-2 and MMP-9. In this work, we present a unified cheminformatics machine-learning framework for predicting potent small-molecule inhibitors of CYP19A1, MMP-2, and MMP-9 using curated ChEMBL bioactivity data. Experimentally measured IC50 records were filtered, standardized, and deduplicated; compounds were labeled using fixed potency thresholds. Each molecule was encoded using a hybrid structural representation that concatenates ECFP4 fingerprints with MACCS keys, capturing complementary topological and predefined substructure information. We benchmarked six modeling strategies Random Forest, LightGBM, Support Vector Classification with a Tanimoto kernel, k-Nearest Neighbors with Tanimoto distance, L2-regularized Logistic Regression, and a probability-level blended ensemble under stratified $k$-fold cross-validation, reporting ROC--AUC, average precision, accuracy, and F1-score. Across targets, the hybrid-fingerprint pipelines achieved strong discrimination and early-enrichment capability, with best ROC--AUC values of 0.948 (CYP19A1), 0.972 (MMP-2), and 0.965 (MMP-9), alongside consistently high PR--AUC.

Source: Computer Methods and Programs in Biomedicine Update

DOI: 10.1016/j.cmpbup.2026.100254

Authors: Bilal Ahmad, Khmaies Ouahada, Farah Jabeen, Habib Hamam

Institutions: University of Johannesburg, Université de Moncton

Language: English

Topic: Computational Drug Discovery Methods

Subfield: Computational Theory and Mathematics

Field: Computer Science

Domain: Physical Sciences

SDG: Good health and well-being

Open Access status: gold

Estimated APC paid: $1,870