English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Although substantial efforts have been made using graph neural networks(GNNs) for AI-driven drug discovery (AIDD), effective molecular representationlearning remains an open challenge, especially in the case of insufficientlabeled molecules. Recent studies suggest that big GNN models pre-trained byself-supervised learning on unlabeled datasets enable better transferperformance in downstream molecular property prediction tasks. However, theapproaches in these studies require multiple complex self-supervised tasks andlarge-scale datasets, which are time-consuming, computationally expensive, anddifficult to pre-train end-to-end. Here, we design a simple yet effectiveself-supervised strategy to simultaneously learn local and global informationabout molecules, and further propose a novel bi-branch masked graph transformerautoencoder (BatmanNet) to learn molecular representations. BatmanNet featurestwo tailored complementary and asymmetric graph autoencoders to reconstruct themissing nodes and edges, respectively, from a masked molecular graph. With thisdesign, BatmanNet can effectively capture the underlying structure and semanticinformation of molecules, thus improving the performance of molecularrepresentation. BatmanNet achieves state-of-the-art results for multiple drugdiscovery tasks, including molecular properties prediction, drug-druginteraction, and drug-target interaction, on 13 benchmark datasets,demonstrating its great potential and superiority in molecular representationlearning.

BatmanNet: Bi-branch Masked Graph Transformer Autoencoder for Molecular Representation