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

Virtual screening plays a critical role in modern drug discovery by enablingthe identification of promising candidate molecules for experimentalvalidation. Traditional machine learning methods such, as Support VectorMachines (SVM) and XGBoost, rely on predefined molecular representations, oftenleading to information loss and potential bias. In contrast, deep learningapproaches-particularly Graph Convolutional Networks (GCNs)-offer a moreexpressive and unbiased alternative by operating directly on molecular graphs.Meanwhile, Large Language Models (LLMs) have recently demonstratedstate-of-the-art performance in drug design, thanks to their capacity tocapture complex chemical patterns from large-scale data via attentionmechanisms.  In this paper, we propose a hybrid architecture that integrates GCNs withLLM-derived embeddings to combine localized structural learning with globalchemical knowledge. The LLM embeddings can be precomputed and stored in amolecular feature library, removing the need to rerun the LLM during trainingor inference and thus maintaining computational efficiency. We found thatconcatenating the LLM embeddings after each GCN layer-rather than only at thefinal layer-significantly improves performance, enabling deeper integration ofglobal context throughout the network. The resulting model achieves superiorresults, with an F1-score of (88.8\%), outperforming standalone GCN (87.9%),XGBoost (85.5%), and SVM (85.4%) baselines.

Combining GCN Structural Learning with LLM Chemical Knowledge for Enhanced Virtual Screening