Machine Learning-Driven De Novo Design of DDP-4 Targeted Drug Molecules Using RNN-Based Generative Models and Reinforcement Learning

Jiatong Lou

doi:10.54254/2753-8818/2025.AU28436

Theoretical and Natural ScienceOpen access

Machine Learning-Driven De Novo Design of DDP-4 Targeted Drug Molecules Using RNN-Based Generative Models and Reinforcement Learning

Research Article

Open Access

Machine Learning-Driven De Novo Design of DDP-4 Targeted Drug Molecules Using RNN-Based Generative Models and Reinforcement Learning

Jiatong Lou ^1*

¹ Cushing Academy

^*Corresponding author: jasminelou0112@gmail.com

Published on 23 October 2025

TNS Vol.144

ISSN (Print): 2753-8826

ISSN (Online): 2753-8818

ISBN (Print): 978-1-80590-441-0

ISBN (Online): 978-1-80590-442-7

Download Cover

Abstract

We present a fully in silico pipeline for the de novo design of dipeptidyl peptidase-4 (DPP- 4) inhibitors that integrates data-driven curation, transfer learning, and reinforcement learning (RL) within the REINVENT architecture. Activity records for human DPP-4 (CHEMBL284) were programmatically retrieved from ChEMBL, normalized to nanomolar units, filtered at IC50≤100 nM, standardized to canonical SMILES, and consolidated into a high-quality training table; pIC50 values were computed and a top-100 reference set was exported for down- stream novelty control. A REINVENT prior was adapted to the DPP-4 chemical space via maximum-likelihood fine-tuning on 173 nonredundant, high-activity SMILES. The adapted generator was then optimized with an RL objective that combined predicted potency (pIC50), drug-likeness (QED), synthetic accessibility (SA), and novelty penalties relative to the top-100 reference inhibitors. Relative to the transfer-learned baseline, RL increased mean QED by ~ 10%, improved normalized synthetic accessibility (1 - SA)/10 by ~ 15%, and maintained diversity with ~60% novelty, while the composite reward showed a clear upward shift. Structure-based evaluation further corroborated these gains: 100 RL-generated molecules achieved a mean docking score of -9.8 kcal/mol, surpassing both pre-RL de novo samples (-7.7 kcal/mol) and the top 100 reference actives (-8.5 kcal/mol). These results demonstrate that RL fine-tuning can steer a pretrained generator toward DPP-4–relevant regions of chemical space with improved developability surrogates and predicted binding. Future work will integrate ADMET predictors into the reward and prioritize wet-lab validation to confirm biochemical potency and advance selected designs toward lead optimization.

Keywords:

DPP-4 inhibitors, de novo drug design, reinforcement learning, generative models, computational drug discovery

View PDF

References

[1]. Istrate, D.; Crisan, L. Dipeptidyl peptidase 4 inhibitors in type 2 diabetes mellitus management: Pharmacophore virtual screening, molecular docking, pharmacokinetic evaluations, and conceptual DfT analysis. Processes 2023, 11, 3100.

[2]. Green, B. D.; Flatt, P. R.; Bailey, C. J. Dipeptidyl peptidase IV (DPP IV) inhibitors: a newly emerging drug class for the treatment of type 2 diabetes. Diabetes and vascular disease re- search 2006, 3, 159–165.

[3]. Petrov, V.; Aleksandrova, T.; Pashev, A. Synthetic Approaches to Novel DPP-IV Inhibitors— A Literature Review. Molecules 2025, 30, 1043.

[4]. Hossain, D.; Saghapour, E.; Chen, J. Y. NeSyDPP4-QSAR: Discovering DPP-4 Inhibitors for Diabetes Treatment with a Neuro-symbolic AI Approach. Frontiers in Bioinformatics 2025, 5, 1603133.

[5]. Olivecrona, M.; Blaschke, T.; Engkvist, O.; Chen, H. Molecular de-novo design through deep reinforcement learning. Journal ofcheminformatics 2017, 9, 48.

[6]. Popova, M.; Isayev, O.; Tropsha, A. Deep reinforcement learning for de novo drug design. Science advances 2018, 4, eaap7885.

[7]. Gaulton, A.; Bellis, L. J.; Bento, A. P.; Chambers, J.; Davies, M.; Hersey, A.; Light, Y.; McGlinchey, S.; Michalovich, D.; Al-Lazikani, B.; Overington, J. P. ChEMBL: a large-scale bioactivity database for drug discovery. Nucleic acids research 2012, 40, D1100–D1107.

[8]. Loeffler, H. H.; He, J.; Tibo, A.; Janet, J. P.; Voronov, A.; Mervin, L. H.; Engkvist, O. Reinvent 4: modern AI–driven generative molecule design. Journal of Cheminformatics 2024, 16, 20.

[9]. Chen, T.; Guestrin, C. Xgboost: A scalable tree boosting system. Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining. 2016; pp 785–794.

[10]. Van De Waterbeemd, H.; Gifford, E. ADMET in silico modelling: towards prediction paradise? Nature reviews Drug discovery 2003, 2, 192-204.

References

[3]. Petrov, V.; Aleksandrova, T.; Pashev, A. Synthetic Approaches to Novel DPP-IV Inhibitors— A Literature Review. Molecules 2025, 30, 1043.

[4]. Hossain, D.; Saghapour, E.; Chen, J. Y. NeSyDPP4-QSAR: Discovering DPP-4 Inhibitors for Diabetes Treatment with a Neuro-symbolic AI Approach. Frontiers in Bioinformatics 2025, 5, 1603133.

[5]. Olivecrona, M.; Blaschke, T.; Engkvist, O.; Chen, H. Molecular de-novo design through deep reinforcement learning. Journal ofcheminformatics 2017, 9, 48.

[6]. Popova, M.; Isayev, O.; Tropsha, A. Deep reinforcement learning for de novo drug design. Science advances 2018, 4, eaap7885.

[9]. Chen, T.; Guestrin, C. Xgboost: A scalable tree boosting system. Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining. 2016; pp 785–794.

[10]. Van De Waterbeemd, H.; Gifford, E. ADMET in silico modelling: towards prediction paradise? Nature reviews Drug discovery 2003, 2, 192-204.

Cite this article

Lou,J. (2025). Machine Learning-Driven De Novo Design of DDP-4 Targeted Drug Molecules Using RNN-Based Generative Models and Reinforcement Learning. Theoretical and Natural Science,144,44-50.

Data availability

The datasets used and/or analyzed during the current study will be available from the authors upon reasonable request.

About volume

Volume title: Proceedings of ICBioMed 2025 Symposium: AI for Healthcare: Advanced Medical Data Analytics and Smart Rehabilitation

ISBN: 978-1-80590-441-0(Print) / 978-1-80590-442-7(Online)

Editor: Alan Wang

Conference website: https://2025.icbiomed.org/auckland.html

Conference date: 17 October 2025

Series: Theoretical and Natural Science

Volume number: Vol.144

ISSN: 2753-8818(Print) / 2753-8826(Online)

© 2024 by the author(s). Licensee EWA Publishing, Oxford, UK. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license. Authors who publish this series agree to the following terms:

1. Authors retain copyright and grant the series right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this series.

2. Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the series's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this series.

3. Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See Open access policy for details).