Clarke Neil D, Ezkurdia Iakes, Kopp Jürgen, Read Randy J, Schwede Torsten, Tress Michael
Genome Institute of Singapore, 138627 Singapore.
Proteins. 2007;69 Suppl 8:10-8. doi: 10.1002/prot.21686.
Experimentally determined protein structures formed the basis of the CASP7 prediction assessments. These target structures were assigned to one or more tertiary structure prediction categories and where necessary were divided into structural domains. Boundaries for these domains were based on visual inspection of the targets and superpositions of the target with template structures. Target domains were classified into three different categories for assessment: "high accuracy modeling," "template-based modeling," and "free modeling." Assessment categories were determined by structural similarity between the target domain and the nearest structural templates in the PDB and by the accuracy of the models submitted by the predictors or by whether or not template information was used to generate the predictions. In CASP7 108 of the 123 target domains were evaluated in the template-based modeling category and the remaining 15 target domains were classified as free modeling. A total of 28 target domains from the template-based modeling category were also assessed in the high accuracy category and four overlapped with the free modeling category.
通过实验确定的蛋白质结构构成了CASP7预测评估的基础。这些目标结构被归入一个或多个三级结构预测类别，并在必要时被划分为结构域。这些结构域的边界基于对目标的目视检查以及目标与模板结构的叠加。目标结构域被分为三类进行评估：“高精度建模”、“基于模板的建模”和“自由建模”。评估类别由目标结构域与蛋白质数据库（PDB）中最接近的结构模板之间的结构相似性以及预测者提交模型的准确性或是否使用模板信息来生成预测来确定。在CASP7中，123个目标结构域中的108个在基于模板的建模类别中进行评估，其余15个目标结构域被归类为自由建模。基于模板的建模类别中的总共28个目标结构域也在高精度类别中进行评估，其中四个与自由建模类别重叠。
