Kolinski A, Skolnick J
Department of Molecular Biology, Scripps Research Institute, La Jolla, California 92037.
Proteins. 1994 Apr;18(4):338-52. doi: 10.1002/prot.340180405.
A new hierarchical method for the simulation of the protein folding process and the de novo prediction of protein three-dimensional structure is proposed. The reduced representation of the protein alpha-carbon backbone employs lattice discretizations of increasing geometrical resolution and a single ball representation of side chain rotamers. In particular, coarser and finer lattice backbone descriptions are used. The coarser (finer) lattice represents C alpha traces of native proteins with an accuracy of 1.0 (0.7) A rms. Folding is simulated by means of very fast Monte Carlo lattice dynamics. The potential of mean force, predominantly of statistical origin, contains several novel terms that facilitate the cooperative assembly of secondary structure elements and the cooperative packing of the side chains. Particular contributions to the interaction scheme are discussed in detail. In the accompanying paper (Kolinski, A., Skolnick, J. Monte Carlo simulation of protein folding. II. Application to protein A, ROP, and crambin. Proteins 18:353-366, 1994), the method is applied to three small globular proteins.
提出了一种用于模拟蛋白质折叠过程和从头预测蛋白质三维结构的新的分层方法。蛋白质α-碳骨架的简化表示采用了几何分辨率不断提高的晶格离散化以及侧链旋转异构体的单球表示。具体而言，使用了较粗和较细的晶格骨架描述。较粗（较细）的晶格以均方根误差1.0（0.7）埃的精度表示天然蛋白质的Cα轨迹。通过非常快速的蒙特卡罗晶格动力学模拟折叠过程。平均力势主要源于统计，包含几个新项，这些新项有助于二级结构元件的协同组装和侧链的协同堆积。详细讨论了对相互作用方案的具体贡献。在随附的论文（科林斯基，A.，斯科尔尼克，J.蛋白质折叠的蒙特卡罗模拟。II.应用于蛋白质A、ROP和胰凝乳蛋白酶原。蛋白质18:353 - 366，1994）中，该方法应用于三种小的球状蛋白质。
