中文摘要：

在蛋白质相互作用基础上,热点残基有聚集在一起形成模块的倾向,这样的模块被定义为热区。热区对揭示生物体的生命活动起着重要的作用,因此,如何有效而精确地对热区进行预测,是一个重要的研究方向。作者结合热区空间密度普遍大于非热区空间密度的特性,先采用聚类算法得到簇集,而后利用相关计算模型得到的结合自由能值过滤掉簇中非热点残基及某些不符合条件的簇。最后,使用社区探测技术处理过滤后的簇集,从而得到最终的预测热区。实验结果显示该热区预测方法具有较好的预测效果。

英文摘要：

On the basis of protein-protein interaction, hot spots tend to get together to form modules, so these modules are defined as hot regions. Hot region plays an important role in revealing the life activities of organisms. Therefore, how to predict hot regions effectively and accurately is a vital research direction. According to the feature that the spatial density of hot regions is denser than that of non-hot regions, the authors used clustering algorithm to get clusters firstly. Then, they utilized the binding free energy calculated by related calculation model to filter out some non-hotspots in clusters and unqualified clusters. Finally, they applied community detection knowledge to the optimized clusters to get the final hot regions. The experimental results show that this method of predicting hot regions has an good effect.