article · Journal of Chemical Information and Modeling
Virtual screening assists modern drug discovery by identifying potential therapeutic molecules prior to laboratory testing. Standard machine learning models often depend on fixed representations that can lose key chemical details, while graph convolutional networks process molecular structures directly. Concurrently, large language models have shown high efficacy in identifying chemical patterns across broad datasets. To leverage both strengths, a hybrid computational architecture integrates graph convolutional networks with embeddings derived from large language models. Rather than adding these embeddings at the end, the system inserts them progressively after each graph layer, continuously enriching structural data with broad chemical context. Precomputing the language model embeddings keeps computing requirements manageable. In benchmark tests across kinase and non-kinase biological targets, the hybrid architecture achieved 88.7% accuracy, outperforming individual graph networks, transformer baselines, and traditional machine learning methods.
Identifying viable drug candidates computationally saves substantial time and laboratory resources. By capturing both detailed structural shapes and broader chemical context, this approach improves screening accuracy. Higher accuracy reduces false positives, helping research teams prioritise the most promising molecules for physical validation more reliably.
This methodology is aimed at pharmaceutical and biotechnology discovery teams conducting virtual screening campaigns. Tested on benchmark biological datasets, the technology sits at an applied research stage. It could be integrated into computational drug discovery pipelines to speed up candidate prioritisation and lower experimental screening costs.
AI-generated from the published abstract. Always read the original work before citing.
Virtual screening plays a critical role in modern drug discovery by enabling the identification of promising candidate molecules for experimental validation. Traditional machine learning methods, such as support vector machines (SVM) and XGBoost, rely on predefined molecular representations, often leading to information loss and potential bias. In contrast, deep learning approaches, particularly graph convolutional networks (GCNs), offer a more expressive and unbiased alternative by operating directly on molecular graphs. Meanwhile, large language models (LLMs) have recently demonstrated state-of-the-art performance in drug design thanks to their capacity to capture complex chemical patterns from large-scale data via attention mechanisms. In this paper, we propose a novel hybrid architecture that combines GCNs with LLM-derived embeddings, evaluated on both kinase-related data sets, which are well-established therapeutic targets of high biological significance, and non-kinase data sets such as the glucocorticoid receptor and PPARG, demonstrating the broader applicability of our approach. Our model introduces a layer-wise concatenation strategy, where LLM embeddings are injected after each GCN layer rather than solely at the final layer. This design enables progressive enrichment of the learned molecular representations with global chemical context throughout the network's depth. The LLM embeddings can be precomputed and stored in a molecular feature library, maintaining computational efficiency during training and inference. We conduct a comprehensive comparison against standalone GCN, Molformer, SVM, and XGBoost baselines, demonstrating that our method achieves superior performance, with an accuracy of 88.7%, compared to 86.8% for GCN, 85.1 for molformer, 85.0% for XGBoost, and 84.7% for SVM. These improvements are practically significant in real-world virtual screening scenarios, where even small gains can reduce false positives and accelerate candidate prioritization.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.1021/acs.jcim.5c00950
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
