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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Molecular property prediction refers to the task of labeling molecules withsome biochemical properties, playing a pivotal role in the drug discovery anddesign process. Recently, with the advancement of machine learning, deeplearning-based molecular property prediction has emerged as a solution to theresource-intensive nature of traditional methods, garnering significantattention. Among them, molecular representation learning is the key factor formolecular property prediction performance. And there are lots ofsequence-based, graph-based, and geometry-based methods that have beenproposed. However, the majority of existing studies focus solely on onemodality for learning molecular representations, failing to comprehensivelycapture molecular characteristics and information. In this paper, a novelmulti-modal representation learning model, which integrates the sequence,graph, and geometry characteristics, is proposed for molecular propertyprediction, called SGGRL. Specifically, we design a fusion layer to fusion therepresentation of different modalities. Furthermore, to ensure consistencyacross modalities, SGGRL is trained to maximize the similarity ofrepresentations for the same molecule while minimizing similarity for differentmolecules. To verify the effectiveness of SGGRL, seven molecular datasets, andseveral baselines are used for evaluation and comparison. The experimentalresults demonstrate that SGGRL consistently outperforms the baselines in mostcases. This further underscores the capability of SGGRL to comprehensivelycapture molecular information. Overall, the proposed SGGRL model showcases itspotential to revolutionize molecular property prediction by leveragingmulti-modal representation learning to extract diverse and comprehensivemolecular insights. Our code is released athttps://github.com/Vencent-Won/SGGRL.

Multi-Modal Representation Learning for Molecular Property Prediction: Sequence, Graph, Geometry