Prodrug-ML - Prodrug-likeness prediction via machine learning on sampled negative decoys

Aug 5, 2026

A prodrug is a pharmacologically inactive derivative that undergoes bioreversible transformation in vivo to release an active parent drug, improving solubility, permeability, and targeting. Prodrug-ML addresses the lack of reliable negative examples in in silico prodrug screening that leads to class imbalance and benchmark unreliability. The method prioritizes prodrug-likeness candidates and reduces wet-lab burden by scoring candidate structures, using property-controlled negative cohorts and decoy hardening, and delivering strong discrimination and early retrieval metrics.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages47

Reading Time1 h 58 min

AccessFree to download

Abstract

Prodrug concept and screening challenge

Prodrug-ML approach and workflow

Negative cohort construction and safeguards

Evaluation protocol and benchmark results

Introduction

Origins of the prodrug concept

University of Birmingham Prodrug-ML Ugurlu, Sadettin Y. ; He, Shan DOI: 10.1007/s10822-025-00725-x License: Creative Commons: Attribution (CC BY) Document Version Publisher's PDF, also known as Version of record Citation for published version (Harvard): Ugurlu, SY & He, S 2026, 'Prodrug-ML: prodrug-likeness prediction via machine learning on sampled negative decoys', Journal of Computer-Aided Molecular Design, vol. 40, no. 1, 45. [https://doi.org/10.1007/s10822-025-](https://doi.org/10.1007/s10822-025-)[ ](https://doi.org/10.1007/s10822-025-)[00725-x](00725-x) Link to publication on Research at Birmingham portal General rights Unless a licence is specified above, all rights (including copyright and moral rights) in this document are retained by the authors and/or the copyright holders. The express permission of the copyright holder must be obtained for any use of this material other than for purposes permitted by law. •Users may freely distribute the URL that is used to identify this publication. •Users may download and/or print one copy of the publication from the University of Birmingham research portal for the purpose of private study or non-commercial research. •User may use extracts from the document in line with the concept of ‘fair dealing’ under the Copyright, Designs and Patents Act 1988 (?) •Users may not further distribute the material nor use it for the purposes of commercial gain. Where a licence is displayed above, please note the terms and conditions of the licence govern your use of this document. When citing, please reference the published version. Take down policy While the University of Birmingham exercises care and attention in making items available there are rare occasions when an item has been uploaded in error or has been deemed to be commercially or otherwise sensitive. If you believe that this is the case for this document, [please contact UBIRA@lists.bham.ac.uk](please contact UBIRA@lists.bham.ac.uk) providing details and we will remove access to the work immediately and investigate. Download date: 26. Feb. 2026 Prodrug-ML: prodrug-likeness prediction via machine learning on sampled negative decoys SadettinY. Ugurlu1,2 · Shan He3 Received: 22 September 2025 / Accepted: 23 November 2025 © The Author(s) 2026 Abstract A prodrug is a pharmacologically inactive (or attenuated) derivative that undergoes bioreversible transformation in vivo to release an active parent drug, enabling temporary optimization of properties such as solubility, permeability, and targeting. Despite expanding catalogs of known prodrugs, in silico screening remains limited by the absence of reliable negative examples: training/evaluation sets often contain only positives or ad-hoc decoys, leading to class imbalance, propertymismatch shortcuts, and irreproducible benchmarks. Unfortunately, the limitation of reliable negatives has resulted in there being no efficient machine learning-based prodrug screening approach. Therefore, we introduce Prodrug-ML, an efficient machine learning-based screen for prodrug-likeness that prioritizes candidates rather than asserting mechanistic truth. Prodrug-ML helps medicinal chemists triage prodrugging ideas during hit-to-lead and lead optimization, filter enumerated libraries of promoiety–attachment variants before ADMET assays, and retrospectively mine internal/ChEMBL-like collections to surface likely prodrug chemotypes. In practice, users (i) generate or collect candidate structures (e.g., parent drug ± pro-moieties), (ii) score them with Prodrug-ML, and (iii) advance only high-scoring candidates to synthesis/assay, thereby reducing wet-lab load while maintaining chemical diversity. In order to achieve such practical usage, the ProdrugML framework, containing the default classifier, LightGBM, addresses these issues by (i) constructing three complementary, property-controlled negative cohorts (DUD-E–style near-misses, random ChEMBL, and stric

FAQs

What problem does Prodrug-ML target in prodrug screening?

In silico screening is limited by unreliable negative examples, which creates class imbalance, property-mismatch shortcuts, and irreproducible benchmarks. Prodrug-ML focuses on ranking prodrug-likeness candidates despite this constraint.

How does Prodrug-ML generate training and evaluation negatives?

It constructs three complementary, property-controlled negative cohorts, including DUD-E–style near-misses, random ChEMBL samples, and strictly filtered ChEMBL. It also applies hardness control and label-noise guardrails for decoys.

How does Prodrug-ML help medicinal chemists in practice?

Users generate or collect candidate structures, score them with Prodrug-ML, and advance only high-scoring candidates to synthesis and assay. This reduces wet-lab load while maintaining chemical diversity during hit-to-lead and lead optimization.

Aug 5, 2026

Upload by Valentina