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

Learning effective protein representations is critical in a variety of tasksin biology such as predicting protein function or structure. Existingapproaches usually pretrain protein language models on a large number ofunlabeled amino acid sequences and then finetune the models with some labeleddata in downstream tasks. Despite the effectiveness of sequence-basedapproaches, the power of pretraining on known protein structures, which areavailable in smaller numbers only, has not been explored for protein propertyprediction, though protein structures are known to be determinants of proteinfunction. In this paper, we propose to pretrain protein representationsaccording to their 3D structures. We first present a simple yet effectiveencoder to learn the geometric features of a protein. We pretrain the proteingraph encoder by leveraging multiview contrastive learning and differentself-prediction tasks. Experimental results on both function prediction andfold classification tasks show that our proposed pretraining methods outperformor are on par with the state-of-the-art sequence-based methods, while usingmuch less pretraining data. Our implementation is available athttps://github.com/DeepGraphLearning/GearNet.

Protein Representation Learning by Geometric Structure Pretraining