DE NOVO MOLECULE GENERATION WITH GRAPH LATENT DIFFUSION MODEL

Conghao Wang; Hiok Hian Ong; Shunsuke Chiba; Jagath C. Rajapakse

doi:10.1109/ICASSP48485.2024.10447480

DE NOVO MOLECULE GENERATION WITH GRAPH LATENT DIFFUSION MODEL

Conghao Wang, Hiok Hian Ong, Shunsuke Chiba, Jagath C. Rajapakse^*

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

De novo generation of molecules is a crucial task in drug discovery. The blossom of deep learning-based generative models, especially diffusion models, has brought forth promising advancements in de novo drug design by finding optimal molecules in a directed manner. However, due to the complexity of chemical space, existing approaches can only generate extremely small molecules. In this study, we propose a Graph Latent Diffusion Model (GLDM) that operates a diffusion model in the latent space modeled by a pretrained autoencoder. Applying diffusion processs on latent representations rather than original molecular graphs, GLDM improves training efficiency and enables generation of larger drug-like molecules. GLDM achieves state-of-the-art results on GuacaMol benchmarks.

Original language	English
Title of host publication	2024 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2024 - Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	2121-2125
Number of pages	5
ISBN (Electronic)	9798350344851
DOIs	https://doi.org/10.1109/ICASSP48485.2024.10447480
Publication status	Published - 2024
Externally published	Yes
Event	49th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2024 - Seoul, Korea, Republic of Duration: Apr 14 2024 → Apr 19 2024

Publication series

Name	ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
ISSN (Print)	1520-6149

Conference

Conference	49th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2024
Country/Territory	Korea, Republic of
City	Seoul
Period	4/14/24 → 4/19/24

Bibliographical note

Publisher Copyright:
© 2024 IEEE.

ASJC Scopus Subject Areas

Software
Signal Processing
Electrical and Electronic Engineering

Keywords

deep learning
drug design
Latent diffusion model
molecule generation

Access to Document

10.1109/ICASSP48485.2024.10447480

Cite this

Wang, C., Ong, H. H., Chiba, S., & Rajapakse, J. C. (2024). DE NOVO MOLECULE GENERATION WITH GRAPH LATENT DIFFUSION MODEL. In 2024 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2024 - Proceedings (pp. 2121-2125). (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICASSP48485.2024.10447480

Wang, Conghao ; Ong, Hiok Hian ; Chiba, Shunsuke et al. / DE NOVO MOLECULE GENERATION WITH GRAPH LATENT DIFFUSION MODEL. 2024 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2024 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2024. pp. 2121-2125 (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings).

@inproceedings{5fe34e5dabae433b8c41f509b384fba0,

title = "DE NOVO MOLECULE GENERATION WITH GRAPH LATENT DIFFUSION MODEL",

abstract = "De novo generation of molecules is a crucial task in drug discovery. The blossom of deep learning-based generative models, especially diffusion models, has brought forth promising advancements in de novo drug design by finding optimal molecules in a directed manner. However, due to the complexity of chemical space, existing approaches can only generate extremely small molecules. In this study, we propose a Graph Latent Diffusion Model (GLDM) that operates a diffusion model in the latent space modeled by a pretrained autoencoder. Applying diffusion processs on latent representations rather than original molecular graphs, GLDM improves training efficiency and enables generation of larger drug-like molecules. GLDM achieves state-of-the-art results on GuacaMol benchmarks.",

keywords = "deep learning, drug design, Latent diffusion model, molecule generation",

author = "Conghao Wang and Ong, {Hiok Hian} and Shunsuke Chiba and Rajapakse, {Jagath C.}",

note = "Publisher Copyright: {\textcopyright} 2024 IEEE.; 49th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2024 ; Conference date: 14-04-2024 Through 19-04-2024",

year = "2024",

doi = "10.1109/ICASSP48485.2024.10447480",

language = "English",

series = "ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "2121--2125",

booktitle = "2024 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2024 - Proceedings",

address = "United States",

}

Wang, C, Ong, HH, Chiba, S & Rajapakse, JC 2024, DE NOVO MOLECULE GENERATION WITH GRAPH LATENT DIFFUSION MODEL. in 2024 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2024 - Proceedings. ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings, Institute of Electrical and Electronics Engineers Inc., pp. 2121-2125, 49th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2024, Seoul, Korea, Republic of, 4/14/24. https://doi.org/10.1109/ICASSP48485.2024.10447480

DE NOVO MOLECULE GENERATION WITH GRAPH LATENT DIFFUSION MODEL. / Wang, Conghao; Ong, Hiok Hian; Chiba, Shunsuke et al.
2024 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2024 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2024. p. 2121-2125 (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - DE NOVO MOLECULE GENERATION WITH GRAPH LATENT DIFFUSION MODEL

AU - Wang, Conghao

AU - Ong, Hiok Hian

AU - Chiba, Shunsuke

AU - Rajapakse, Jagath C.

PY - 2024

Y1 - 2024

N2 - De novo generation of molecules is a crucial task in drug discovery. The blossom of deep learning-based generative models, especially diffusion models, has brought forth promising advancements in de novo drug design by finding optimal molecules in a directed manner. However, due to the complexity of chemical space, existing approaches can only generate extremely small molecules. In this study, we propose a Graph Latent Diffusion Model (GLDM) that operates a diffusion model in the latent space modeled by a pretrained autoencoder. Applying diffusion processs on latent representations rather than original molecular graphs, GLDM improves training efficiency and enables generation of larger drug-like molecules. GLDM achieves state-of-the-art results on GuacaMol benchmarks.

AB - De novo generation of molecules is a crucial task in drug discovery. The blossom of deep learning-based generative models, especially diffusion models, has brought forth promising advancements in de novo drug design by finding optimal molecules in a directed manner. However, due to the complexity of chemical space, existing approaches can only generate extremely small molecules. In this study, we propose a Graph Latent Diffusion Model (GLDM) that operates a diffusion model in the latent space modeled by a pretrained autoencoder. Applying diffusion processs on latent representations rather than original molecular graphs, GLDM improves training efficiency and enables generation of larger drug-like molecules. GLDM achieves state-of-the-art results on GuacaMol benchmarks.

KW - deep learning

KW - drug design

KW - Latent diffusion model

KW - molecule generation

UR - http://www.scopus.com/inward/record.url?scp=85195427999&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85195427999&partnerID=8YFLogxK

U2 - 10.1109/ICASSP48485.2024.10447480

DO - 10.1109/ICASSP48485.2024.10447480

M3 - Conference contribution

AN - SCOPUS:85195427999

T3 - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

SP - 2121

EP - 2125

BT - 2024 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2024 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 49th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2024

Y2 - 14 April 2024 through 19 April 2024

ER -

Wang C, Ong HH, Chiba S, Rajapakse JC. DE NOVO MOLECULE GENERATION WITH GRAPH LATENT DIFFUSION MODEL. In 2024 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2024 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2024. p. 2121-2125. (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings). doi: 10.1109/ICASSP48485.2024.10447480

DE NOVO MOLECULE GENERATION WITH GRAPH LATENT DIFFUSION MODEL

Abstract

Publication series

Conference

Bibliographical note

ASJC Scopus Subject Areas

Keywords

Access to Document

Other files and links

Fingerprint

Cite this