基于距离特征的蛋白质无序段建模方法及应用

Protein disordered region modeling based on distance and its application

蛋白质无序段在天然条件下无法折叠到特定的三维结构,具有高度的结构动态性。现有的对无序段从能量角度建模的方法缺少对无序段的空间约束。首先提取无序段的空间特征,并找到使预测过程趋于最低能量的最优特征值,最后将这些特征值作为约束应用到并行蚁群蛋白质结构预测算法。实验结果表明:该方法可以有效地利用无序段对蛋白质结构的连接作用,有助于搜索到更接近天然结构的折叠区域结构。

Proteins contain intrinsically disordered regions. Under the natural conditions, the disordered region might fail to adopt a stable three-dimensional structure; it exists as a dynamic ensemble. The existing methods, which model disordered regions based on energy function, don＇t show enough spatial constraints. This paper pro- poses to obtain distance characteristics of disordered region,then finds eigenvalues guiding the sampler to ap- proach lowest energy structure. Finally these eigenvalues are employed by parallel ant colonies as spatial con- straints for backbone structure prediction of folded part. The results indicate that our approach can make use of connection features of intrinsically disordered region, and find nearer native conformations of protein folded part.