Dipartimento di Scienze Chimiche, Università di Padova , via Marzolo, 1, I-35131 Padova, Italy.
Département de Chimie, Ecole Normale Supérieure, PSL Research University, UPMC Université Paris 06, CNRS, Laboratoire des Biomolécules (LBM) , 24 rue Lhomond, 75005 Paris, France.
J Chem Theory Comput. 2017 Jan 10;13(1):309-319. doi: 10.1021/acs.jctc.6b00702. Epub 2016 Dec 6.
In this article, we present a clustering method of atoms in proteins based on the analysis of the correlation times of interatomic distance correlation functions computed from MD simulations. The goal is to provide a coarse-grained description of the protein in terms of fewer elements that can be treated as dynamically independent subunits. Importantly, this domain decomposition method does not take into account structural properties of the protein. Instead, the clustering of protein residues in terms of networks of dynamically correlated domains is defined on the basis of the effective correlation times of the pair distance correlation functions. For these properties, our method stands as a complementary analysis to the customary protein decomposition in terms of quasi-rigid, structure-based domains. Results obtained for a prototypal protein structure illustrate the approach proposed.
在本文中，我们提出了一种基于分析 MD 模拟计算的原子间距离相关函数的相关时间的蛋白质中原子的聚类方法。目标是根据较少的元素来提供蛋白质的粗粒度描述，这些元素可以被视为动态独立的亚基。重要的是，这种域分解方法不考虑蛋白质的结构特性。相反，基于对配对距离相关函数的有效相关时间，根据动态相关域的网络对蛋白质残基进行聚类。对于这些特性，我们的方法是对基于准刚性结构域的常规蛋白质分解的补充分析。针对原型蛋白质结构获得的结果说明了所提出的方法。
