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

Robust and efficient interpretation of QSAR methods is quite useful tovalidate AI prediction rationales with subjective opinion (chemist or biologistexpertise), understand sophisticated chemical or biological process mechanisms,and provide heuristic ideas for structure optimization in pharmaceuticalindustry. For this purpose, we construct a multi-layer self-attention basedGraph Neural Network framework, namely Ligandformer, for predicting compoundproperty with interpretation. Ligandformer integrates attention maps oncompound structure from different network blocks. The integrated attention mapreflects the machine's local interest on compound structure, and indicates therelationship between predicted compound property and its structure. This workmainly contributes to three aspects: 1. Ligandformer directly opens theblack-box of deep learning methods, providing local prediction rationales onchemical structures. 2. Ligandformer gives robust prediction in differentexperimental rounds, overcoming the ubiquitous prediction instability of deeplearning methods. 3. Ligandformer can be generalized to predict differentchemical or biological properties with high performance. Furthermore,Ligandformer can simultaneously output specific property score and visibleattention map on structure, which can support researchers to investigatechemical or biological property and optimize structure efficiently. Ourframework outperforms over counterparts in terms of accuracy, robustness andgeneralization, and can be applied in complex system study.

Ligandformer: A Graph Neural Network for Predicting Compound Property with Robust Interpretation