PuSH - Publikationsserver des Helmholtz Zentrums München

Li, Y.* ; Hu, Q.* ; Han, S.* ; Wang-Sattler, R. ; Du, W.*

Multi-manifolds fusing hyperbolic graph network balanced by pareto optimization for identifying spatial domains of spatial transcriptomics.

Brief. Bioinform. 26:bbaf162 (2025)
Verlagsversion DOI PMC
Free by publisher
Creative Commons Lizenzvertrag
Identifying spatial domains for spatial transcriptomics is crucial for achieving comprehensive insights into the pathogenesis of gene expression. Increasingly, computational methods based on graph neural networks are being developed for spatial transcriptomics. However, previous methods have solely focused on the Euclidean manifold. To effectively exploit and explore the informative and deeper topological structures of inherent manifolds, we presented a Multi-Manifolds fusing hyperbolic graph network, balanced by Pareto optimization, for identifying spatial domains in Spatial Transcriptomics (MManiST). First, we developed multi-manifolds encoders for distinct manifolds using the hyperbolic neural network. Features from different manifolds were then combined using an attention mechanism, with multiple reconstruction losses balanced by Pareto optimization. Extensive experiments on commonly used benchmark datasets show that our method consistently outperforms seven state-of-the-art methods. Additionally, we investigated the validity of each component and the impact of fusion methods in ablation experiments.
Altmetric
Weitere Metriken?
[➜Einloggen]
Zusatzinfos bearbeiten [➜Einloggen]
Publikationstyp Artikel: Journalartikel
Dokumenttyp Wissenschaftlicher Artikel
Korrespondenzautor
Schlagwörter Graph Neural Network ; Hyperbolic Space ; Spatial Domain Identification ; Spatial Transcriptomics
ISSN (print) / ISBN 1467-5463
e-ISSN 1477-4054
Zeitschrift Briefings in Bioinformatics
Quellenangaben Band: 26, Heft: 2, Seiten: , Artikelnummer: bbaf162 Supplement: ,
Verlag Oxford University Press
Verlagsort Great Clarendon St, Oxford Ox2 6dp, England
Nichtpatentliteratur Publikationen
Begutachtungsstatus Peer reviewed
Institut(e) Institute of Translational Genomics (ITG)
Förderungen Natural Science Foundation of Jilin Province
National Natural Science Foundation of China