**中国科学院团队突破糖蛋白质作用位点预测难题：深度学习模型DeepGlycanSite准确率超越传统方法30倍**
近日，中国科学院的一支研究团队取得重大突破，他们开发的深度学习模型DeepGlycanSite能够准确预测蛋白质结构上的糖结合位点。该技术的出现，为解决生物学领域中长期存在的难题提供了有效工具。
糖类在自然界中极为丰富，对生命过程至关重要。了解糖与蛋白质间的相互作用对于揭示生物学机理和开发新疗法至关重要。但糖分子的复杂性和多样性使得实验识别糖-蛋白质结合位点变得困难重重。
DeepGlycanSite模型结合了蛋白质的几何和进化特征，融入了具有Transformer架构的深度等变图神经网络中。其预测准确率大大超越了传统方法和先进的前置技术。此外，结合诱变研究，该模型还成功揭示了重要G蛋白偶联受体的特定糖类识别位点。
该研究成果以“Highly accurate carbohydrate-binding site prediction with DeepGlycanSite”为题，于2024年6月17日在《自然通讯》杂志上发表。业内专家表示，DeepGlycanSite模型的成功应用，为深入探索糖蛋白功能、理解糖类调节机制提供了有力支持，有望为药物研发和新疗法带来革命性进展。
英语如下：
News Title: “CAS DeepGlycanSite Model: Precision Prediction of Glycan-Protein Interaction Sites”
News Content: **Chinese Academy of Sciences Team Breaks Through in Predicting Glycan-Protein Interaction Sites: Deep Learning Model DeepGlycanSite Accuracy Exceeds Traditional Methods by 30 Times**
Recently, a research team from the Chinese Academy of Sciences has made a significant breakthrough. Their developed deep learning model, DeepGlycanSite, can accurately predict carbohydrate binding sites on protein structures. This technology provides an effective tool for solving long-standing problems in the biological field.
Carbohydrates are extremely abundant in nature and crucial to life processes. Understanding the interaction between carbohydrates and proteins is vital to revealing biological mechanisms and developing new therapies. However, the complexity and diversity of carbohydrate molecules make it challenging to experimentally identify carbohydrate-protein binding sites.
The DeepGlycanSite model combines protein’s geometric and evolutionary features and integrates deep equivariant graph neural networks with Transformer architecture. Its prediction accuracy significantly surpasses traditional methods and advanced predecessor technologies. In addition, combined with mutagenesis research, this model has also successfully revealed specific carbohydrate recognition sites on important G-protein-coupled receptors.
The research results were published on June 17, 2024, in the journal Nature Communications with the title “Highly accurate carbohydrate-binding site prediction with DeepGlycanSite”. Experts in the industry indicate that the successful application of the DeepGlycanSite model provides strong support for exploring the functions of glycoproteins, understanding carbohydrate regulatory mechanisms, and is expected to bring revolutionary progress to drug development and new therapies.
【来源】https://www.jiqizhixin.com/articles/2024-06-25-12
Views: 2