基于氨基酸接触能网络的蛋白质结合位点分析

Protein-protein interactions (PPIs) were involved in almost all the biological processes, thus researches on these interactions were quite important for understanding the pathogenesis of disease and the prediction of drug targets. The analysis of protein-protein binding sites is a critical step for PPIs understanding. The amino acids network model become an emerging paradigm in protein science. In this project, we will systematically analyze the characteristics of the binding sites in protein complexes. We will first construct the amino acid contact energy networks (AACENs) and node-weighted AACENs (NWAACENs) models for protein complexes. Next, based on these network model, we will explore the differences features of binding sites between homo-oligomers and hetero-oligomers through computing the network parameters of binding sites in the networks of monomers and oligomers. Then, the characteristics of binding sites of disordered proteins, who could interact with different proteins to form different protein complexes, will be analyzed and the subnetworks’ topological structure of disordered protein in different complexes will be also compared. At last, the features of binding sites in ‘vehicle of transcription’-RNA polymerase that interact with key proteins in transcription will be explored by the above methods. The findings of the project will provide the basic characteristics for binding sites of different complexes, theoretical evidences for protein-protein binding sites personalized prediction and guidance for finding drug targets.

很多重要的生物过程都是在蛋白质-蛋白质相互作用的参与下完成的，研究这种相互作用关系对理解疾病发生原因、预测药物靶点都有重要的指导意义。基于复杂网络的氨基酸网络模型为蛋白质研究提供了新的范式。本项目将系统地对蛋白质复合物中结合位点的网络性质进行分析。首先，我们将建立蛋白质复合物的氨基酸接触能网络和点加权氨基酸接触能网络。然后基于此网络模型，从中心性、连通性、模块化等网络性质着手比较同源寡聚体和异源寡聚体中结合位点的性质。接着研究可与多个蛋白质结合的无序蛋白上结合位点的网络性质并对复合物中无序蛋白对应子网络进行网络层次的比较分析。最后将上述基于网络的方法应用到“转录机器”的结合位点分析中。通过本项目将得到不同类型蛋白质复合物结合位点的规律和网络特征，为蛋白质结合位点的研究提供新的方法和思路，为蛋白质结合位点个性化预测提供理论依据，为寻找新的药物靶点提供一定的指导。

蛋白质在大多数生物功能和过程中起着不可或缺的作用，但它们的大部分综合性和多样性的功能仍然不明确。网络理论越来越多地应用于描述复杂的生物系统，包括蛋白质。在本项目中，我们首先总结评价了用于描述生物系统的一般氨基酸网络（amino acids network）模型，然后给出了这些网络参数作为药物设计拓扑指标的一些最新应用。这些网络模型使依赖变构效应的新药开发成为可能。接着我们提出了一种节点加权的AAN策略，称为点加权氨基酸接触能量网络（node-weighted amino acid contact energy network, NACEN），用于蛋白质功能残基（FRs）的表征和预测。结果表明，用NACEN参数表征比用未加权网络参数表征的更易区分FRs和非FRs。NACEN具有较少的分类特征就可鉴别FR，并为变构调控提供残基水平预测。该策略可方便地应用于其它功能残基的识别。开发了一个用于NACEN构建和分析的R包和web服务器。