Skip to content
DraconiticResearch
Search Sources API
203,234 studies31 sourcesUpdated 2 min ago
Peer-reviewedCC BYNature Communications

Autonomous bioisosteric replacement for multi-property optimization in drug design

  • Hyeongwoo Kim1
  • Seokhyun Moon1
  • Wonho Zhung1,2
  • Shin‐Woo Kim1
  • Jaechang Lim2
  • Woo Youn Kim3,4,5,6,7
7 affiliations
  1. 1Department of Chemistry, KAIST, Daejeon, Republic of Korea
  2. 2HITS Inc., Seoul, Republic of Korea
  3. 3Department of AX, KAIST, Daejeon, Republic of Korea. [email protected]
  4. 4Department of Chemistry, KAIST, Daejeon, Republic of Korea. [email protected]
  5. 5HITS Inc., Seoul, Republic of Korea. [email protected]
  6. 6InnoCORE AI-CRED Institute, KAIST, Daejeon, Republic of Korea. [email protected]
  7. 7Kim Jaechul Graduate School of AI, KAIST, Seoul, Republic of Korea. [email protected]
Open at sourcePDFDOICite

Abstract

Optimizing molecular properties while preserving biological activity is a central challenge in drug design. Bioisosteric replacement, which substitutes a molecular fragment with a chemically or biologically analogous moiety, offers a powerful strategy for fine-tuning properties without disrupting target binding. However, existing in silico approaches often rely on expert-defined modification sites or struggle to modulate multiple molecular properties simultaneously. Here, we present DeepBioisostere, a deep generative model that performs end-to-end bioisosteric replacement by autonomously selecting and substituting molecular fragments to satisfy multiple target properties. The model captures complex relationships across the molecular graph, enabling the optimization of sophisticated properties such as drug-likeness and synthetic accessibility. By learning from experimental bioassay data, DeepBioisostere proposes replacements that maintain biological activities, even generating potential bioisosteres beyond the training data. We demonstrate the effectiveness of the model in computational hit-to-lead optimization scenarios, highlighting its potential to accelerate rational molecular design without relying on expert heuristics or pre-established substitution rules.

Related

  1. Design and Optimization of Bottom-Hole Temperature–Pressure Combinations in Gas Production from Gas Hydrates via Carbon Dioxide Replacement Strategy
    Jingjuan Wu, Qi Li, Qingchao Li +3
    EnergiesPeer-reviewedCC BYFull text46 citedEnvironmental chemistryOpenAlex
  2. Bioisosteric Design of Novel Flavonoid-Derived Aromatase Inhibitors to Reduce Estradiol Biosynthesis in ER+ Breast Cancer
    Zion Beason, Arslan Sheikh
    ChemRxivPreprintCC BY-NCFull textGeneral chemistryChemRxiv
  3. Autonomous Experimentation Enables Rapid Optimization of Photochemical Continuous-Flow Synthesis of Diazaxanthilidene Fluorophores
    Joao Luiz P. de Albuquerque, Scott Laverty, Mao Niimura +3
    ChemRxivPreprintCC BYFull textGeneral chemistryChemRxiv
  4. Machine learning-aided design of metallic mechanical metamaterials: fabrication-aware modelling, property prediction, and inverse design
    Heyuan Huang, Gang Xue, Zhicheng Dong +4
    Virtual and Physical PrototypingPeer-reviewedCC BY9 citedMaterials & polymersOpenAlex
  5. Polymer Informatics in the Era of Foundation Models: From Representation to Autonomous Design
    Jianxin He, Ying Li
    ChemRxivPreprintCC BY-NC-NDFull textMaterials & polymersChemRxiv

Cite

BibTeXRIS
BibTeX
@article{kim2026autonomous,
  title = {{Autonomous bioisosteric replacement for multi-property optimization in drug design}},
  author = {Hyeongwoo Kim and Seokhyun Moon and Wonho Zhung and Shin‐Woo Kim and Jaechang Lim and Woo Youn Kim},
  journal = {Nature Communications},
  year = {2026},
  month = jul,
  doi = {10.1038/s41467-026-75512-9},
  url = {https://doi.org/10.1038/s41467-026-75512-9},
  note = {CC BY},
}
RIS
TY  - JOUR
TI  - Autonomous bioisosteric replacement for multi-property optimization in drug design
AU  - Hyeongwoo Kim
AU  - Seokhyun Moon
AU  - Wonho Zhung
AU  - Shin‐Woo Kim
AU  - Jaechang Lim
AU  - Woo Youn Kim
PY  - 2026
DA  - 2026/07/27
JO  - Nature Communications
DO  - 10.1038/s41467-026-75512-9
UR  - https://doi.org/10.1038/s41467-026-75512-9
AB  - Optimizing molecular properties while preserving biological activity is a central challenge in drug design. Bioisosteric replacement, which substitutes a molecular fragment with a chemically or biologically analogous moiety, offers a powerful strategy for fine-tuning properties without disrupting target binding. However, existing in silico approaches often rely on expert-defined modification sites or struggle to modulate multiple molecular properties simultaneously. Here, we present DeepBioisostere, a deep generative model that performs end-to-end bioisosteric replacement by autonomously selecting and substituting molecular fragments to satisfy multiple target properties. The model captures complex relationships across the molecular graph, enabling the optimization of sophisticated properties such as drug-likeness and synthetic accessibility. By learning from experimental bioassay data, DeepBioisostere proposes replacements that maintain biological activities, even generating potential bioisosteres beyond the training data. We demonstrate the effectiveness of the model in computational hit-to-lead optimization scenarios, highlighting its potential to accelerate rational molecular design without relying on expert heuristics or pre-established substitution rules.
SN  - 2041-1723
ER  -