Skip to content
DraconiticResearch
Search Sources API
203,234 studies31 sourcesUpdated 37 min ago
Peer-reviewedCC BYNature Communications

Flow matching for reaction pathway generation

  • Ping Tuo1
  • Jiale Chen2
  • Ju Li3,4
4 affiliations
  1. 1Bakar Institute of Digital Materials for the Planet, University of California, Berkeley, Berkeley, CA, USA. [email protected]
  2. 2Institute of Science and Technology Austria, Klosterneuburg, Austria
  3. 3Department of Materials Science and Engineering, Massachusetts Institute of Technology, Cambridge, MA, USA. [email protected]
  4. 4Department of Nuclear Science and Engineering, Massachusetts Institute of Technology, Cambridge, MA, USA. [email protected]
Open at sourcePDFDOICite

Abstract

Elucidating reaction mechanisms requires efficient generation of transition states (TSs) and products. Existing diffusion and sequence-based models accelerate parts of this process over traditional string-based methods, but typically still require manual enumeration of either TSs or products, and stochastic diffusion dynamics can be inefficient and hard to control. We introduce MolGEN, a conditional flow-matching framework that uses deterministic optimal transport to map Gaussian priors to chemical distributions. For TS generation, MolGEN improves TS geometry and barrier-height prediction over diffusion models while enabling sub-second sampling. For reaction product generation, it achieves competitive top-k accuracy while preserving mass and electron balance. Using the same backbone for TS and product sampling, MolGEN enables template-free generative exploration of reaction networks without the repeated quantum-chemistry searches required by prior methods. For the γ-ketohydroperoxide decomposition network, it produces more valid TSs than string-based methods using only 12 quantum-chemistry evaluations instead of 1156, and identifies a lower-barrier pathway.

Related

  1. Symmetry-Aware Flow Matching for End-to-end Molecular Crystal Generation
    Wendi Cai, Fanyang Mo
    arXivPreprintCC BY-NC-ND 4.0Full textPhysical chemistryarXiv
  2. FLOWR.ROOT – A flow matching-based foundation model for joint multi-purpose structure-aware 3D ligand generation and affinity prediction
    Julian Cremer, Tuan Le, Mohammad Mehdi Ghahremanpour +3
    Nature CommunicationsPeer-reviewedCC BYFull text9 citedMedicinal chemistry & drug discoveryOpenAlex
  3. MIRRA: Fast, High-Fidelity Molecular Generation with Chemistry-Informed Discrete Flow Matching
    Ge Sun, Juan J. de Pablo
    arXivPreprintCC BY-NC-SA 4.0Full textPhysical chemistryComputational physicsarXiv
  4. FLOWR: flow matching for structure-aware de novo, interaction- and fragment-based ligand generation
    Julian Cremer, Ross Irwin, Alessandro Tibo +3
    Nature Computational SciencePeer-reviewedCC BYFull text24 citedMedicinal chemistry & drug discoveryOpenAlex
  5. CatWalk: Reaction Pathway Generation and Transition State Sampling via Score-based Diffusion Models
    Teerachote Pakornchote, Nima Karimitari, Jacob Clary +3
    ChemRxivPreprintCC BYFull textMaterials & polymersChemRxiv

Cite

BibTeXRIS
BibTeX
@article{tuo2026flow,
  title = {{Flow matching for reaction pathway generation}},
  author = {Ping Tuo and Jiale Chen and Ju Li},
  journal = {Nature Communications},
  year = {2026},
  month = jul,
  doi = {10.1038/s41467-026-75654-w},
  url = {https://doi.org/10.1038/s41467-026-75654-w},
  note = {CC BY},
}
RIS
TY  - JOUR
TI  - Flow matching for reaction pathway generation
AU  - Ping Tuo
AU  - Jiale Chen
AU  - Ju Li
PY  - 2026
DA  - 2026/07/17
JO  - Nature Communications
DO  - 10.1038/s41467-026-75654-w
UR  - https://doi.org/10.1038/s41467-026-75654-w
AB  - Elucidating reaction mechanisms requires efficient generation of transition states (TSs) and products. Existing diffusion and sequence-based models accelerate parts of this process over traditional string-based methods, but typically still require manual enumeration of either TSs or products, and stochastic diffusion dynamics can be inefficient and hard to control. We introduce MolGEN, a conditional flow-matching framework that uses deterministic optimal transport to map Gaussian priors to chemical distributions. For TS generation, MolGEN improves TS geometry and barrier-height prediction over diffusion models while enabling sub-second sampling. For reaction product generation, it achieves competitive top-k accuracy while preserving mass and electron balance. Using the same backbone for TS and product sampling, MolGEN enables template-free generative exploration of reaction networks without the repeated quantum-chemistry searches required by prior methods. For the γ-ketohydroperoxide decomposition network, it produces more valid TSs than string-based methods using only 12 quantum-chemistry evaluations instead of 1156, and identifies a lower-barrier pathway.
SN  - 2041-1723
ER  -