Skip to content
DraconiticResearch
Search Sources API
194,161 studies31 sourcesUpdated 50 min ago
PreprintCC BY-NCChemRxiv

Multimodal Representation Learning for Exploring Natural Product Chemical Space

  • Tarapong Srisongkram1,2
  • Supreeya Paiboon1,2
  • Huynh Anh Duy3,2,4,5
5 affiliations
  1. 1Division of Pharmaceutical Chemistry
  2. 2Faculty of Pharmaceutical Sciences, Khon Kaen University
  3. 3Graduate School in the Program of Research and Development in Pharmaceuticals
  4. 4Faculty of Health Sciences
  5. 5College of Natural Sciences, Can Tho University
Open at sourceDOICite

Abstract

Natural products represent a major source of therapeutic molecules, yet their structural diversity and limited biological annotations pose substantial challenges for computational exploration and discovery. Here, we develop a multimodal molecular representation framework that integrates chemical language, molecular fingerprints, physicochemical descriptors, and molecular graphs into a unified latent space for natural product discovery. Using structurally out-of-distribution molecular spaces, we show that molecular representation and sampling strategy jointly determine which regions of chemical space are explored. Multimodal pretraining captured chemically informative representations that enabled exploration of structurally unseen natural products, while comparisons with singlemodality representations showed that representation advantages depended on the discovery objective rather than being universally superior. The pretrained representation was transferable across anti-inflammatory and antioxidant activity predictions, and its integration with active learning enabled data-efficient exploration of biologically active molecules. Prospective evaluation of structurally unseen natural products further connected model-guided exploration with experimental antioxidant activity. These findings establish molecular representation and sampling as coupled determinants of chemical-space exploration and show

Related

  1. Scaling multimodal materials representation learning with synthetic narratives
    Yang Jeong Park, Mayank Kumaran, Chia-Wei Hsu +2
    Nature CommunicationsPeer-reviewedCC BY-NC-NDFull text7 citedMaterials & polymersOpenAlex
  2. Combinatorial Steiner Methodology for Efficiently Exploring Chemical Space in CO ₂ Hydrogenation to Fischer-Tropsch Reactivity via High-throughput Experimentation
    Nestor Iwanojko, Erwin Lam, Christophe Copéret
    ChemRxivPreprintCC BYFull textCatalysisChemRxiv
  3. UniPert-G2CP bridges genetic and chemical screens from molecular representation to phenotype modeling
    Yiming Li, Min De ZENG, Jun Zhu +6
    CellPeer-reviewedCC BY9 citedMedicinal chemistry & drug discoveryOpenAlex
  4. Count your bits: fingerprint benchmarking to assess broad chemical space representation
    Florian Huber, Julian Pollmann
    Journal of CheminformaticsPeer-reviewedCC BYFull text7 citedMedicinal chemistry & drug discoveryOpenAlex
  5. Ultrastructural, Chemical, and Wettability Insights into the Interaction between Novel Natural Dentifrices and Nanohybrid Composites: A Multi-Technique Microscopy Approach
    Giovanna dos Santos S. Teixeira, Ilene Cristine R. Cesar, Luís Eduardo S. Soares
    ACS OmegaPeer-reviewedCC BY 4.0Full textGeneral chemistryCrossref

Cite

BibTeXRIS
BibTeX
@misc{srisongkram2026multimodal,
  title = {{Multimodal Representation Learning for Exploring Natural Product Chemical Space}},
  author = {Tarapong Srisongkram and Supreeya Paiboon and Huynh Anh Duy},
  howpublished = {ChemRxiv},
  year = {2026},
  month = sep,
  doi = {10.26434/chemrxiv.15009161/v1},
  url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15009161/v1},
  note = {CC BY-NC},
}
RIS
TY  - UNPB
TI  - Multimodal Representation Learning for Exploring Natural Product Chemical Space
AU  - Tarapong Srisongkram
AU  - Supreeya Paiboon
AU  - Huynh Anh Duy
PY  - 2026
DA  - 2026/09/21
PB  - ChemRxiv
DO  - 10.26434/chemrxiv.15009161/v1
UR  - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15009161/v1
AB  - Natural products represent a major source of therapeutic molecules, yet their structural diversity and limited biological annotations pose substantial challenges for computational exploration and discovery. Here, we develop a multimodal molecular representation framework that integrates chemical language, molecular fingerprints, physicochemical descriptors, and molecular graphs into a unified latent space for natural product discovery. Using structurally out-of-distribution molecular spaces, we show that molecular representation and sampling strategy jointly determine which regions of chemical space are explored. Multimodal pretraining captured chemically informative representations that enabled exploration of structurally unseen natural products, while comparisons with singlemodality representations showed that representation advantages depended on the discovery objective rather than being universally superior. The pretrained representation was transferable across anti-inflammatory and antioxidant activity predictions, and its integration with active learning enabled data-efficient exploration of biologically active molecules. Prospective evaluation of structurally unseen natural products further connected model-guided exploration with experimental antioxidant activity. These findings establish molecular representation and sampling as coupled determinants of chemical-space exploration and show
ER  -