A novel web server predicts amino acid residue protection against hydrogen-deuterium exchange
A novel web server predicts amino acid residue protection against hydrogen-deuterium exchange
复制标题
DOI：
10.1093/bioinformatics/btt168
复制
发表时间：
2013-06-01
期刊：
影响因子：
5.8
通讯作者：
Galzitskaya, Oxana V.
中科院分区：
文献类型：
--
作者：
Lobanov, Mikhail Yu;Suvorina, Masha Yu;Galzitskaya, Oxana V.
动机：为了阐明结构元件和多肽链迁移率之间的关系，需要进行一组结构统计分析。因为目前具有确定空间结构的蛋白质比那些已知氨基酸序列的蛋白质要少得多，所以能够仅根据蛋白质一级结构预测氢-氘（HD）交换的质子保护程度非常重要。结果：在这里，我们提出了一种新颖的网络服务器，旨在仅根据所研究的蛋白质链的一级结构来预测针对HD交换的氨基酸残基保护程度。基于氨基酸序列，所呈现的服务器提供以下三种可能性（预测器）供用户选择。首先，预测该蛋白质中发生的接触数量，这有助于估计防止 HD 交换的质子数量（灵敏度 0.71）。其次，该蛋白质中氢键的概率，这对于查找未受保护的质子的数量很有用（特异性 0.71）。最后是人工预测器的使用。此外，我们还报告了 HD 交换的质谱分析，该分析首次应用于游离氨基酸。其结果与 10 种球状蛋白的理论数据（质子数）非常吻合（相关系数为 0.73）。我们率先编译了 35 种蛋白质的两个实验 HD 交换数据数据集。
Motivation: To clarify the relationship between structural elements and polypeptide chain mobility, a set of statistical analyses of structures is necessary. Because at present proteins with determined spatial structures are much less numerous than those with amino acid sequence known, it is important to be able to predict the extent of proton protection from hydrogen-deuterium (HD) exchange basing solely on the protein primary structure.Results: Here we present a novel web server aimed to predict the degree of amino acid residue protection against HD exchange solely from the primary structure of the protein chain under study. On the basis of the amino acid sequence, the presented server offers the following three possibilities (predictors) for user's choice. First, prediction of the number of contacts occurring in this protein, which is shown to be helpful in estimating the number of protons protected against HD exchange (sensitivity 0.71). Second, probability of H-bonding in this protein, which is useful for finding the number of unprotected protons (specificity 0.71). The last is the use of an artificial predictor. Also, we report on mass spectrometry analysis of HD exchange that has been first applied to free amino acids. Its results showed a good agreement with theoretical data (number of protons) for 10 globular proteins (correlation coefficient 0.73). We pioneered in compiling two datasets of experimental HD exchange data for 35 proteins.