ALGORITHMS FOR THE OPTIMAL IDENTIFICATION OF SEGMENT NEIGHBORHOODS
ALGORITHMS FOR THE OPTIMAL IDENTIFICATION OF SEGMENT NEIGHBORHOODS
复制标题
DOI：
10.1007/bf02458835
复制
发表时间：
1989-01-01
影响因子：
3.5
通讯作者：
LAWRENCE, CE
中科院分区：
文献类型：
--
作者：
AUGER, IE;LAWRENCE, CE
提出了两种有效的分割邻域识别算法。段邻域是一组具有共同特征的连续残基。开发了两个程序来有效地找到描述这些特征的模型参数和定义每个段邻域边界的残差的估计。该算法可以接受几乎任何的分段邻域模型，并且可以应用于包括最小二乘和最大似然在内的广泛的最佳拟合函数。算法依次识别序列中最重要的特征。将其中一种方法应用于影响病毒的血凝素蛋白，揭示了一种可能的构象变化机制，通过发现强七重体结构的断裂。
Two algorithms for the efficient identification of segment neighborhoods are presented. A segment neighborhood is a set of contiguous residues that share common features. Two procedures are developed to efficiently find estimates for the parameters of the model that describe these features and for the residues that define the boundaries of each segment neighborhood. The algorithms can accept nearly any model of segment neighborhood, and can be applied with a broad class of best fit functions including least squares and maximum likelihood. The algorithms successively identify the most important features of the sequence. The application of one of these methods to the haemogglutinin protein of influence virus reveals a possible mechanism for conformational change through the finding of a break in a strong heptad repeat structure.