Johansen Morten Bo, Kiemer Lars, Brunak Søren
Center for Biological Sequence Analysis, BioCentrum-DTU, Technical University of Denmark, Building 208, DK-2800 Lyngby, Denmark.
Glycobiology. 2006 Sep;16(9):844-53. doi: 10.1093/glycob/cwl009. Epub 2006 Jun 8.
Glycation is a nonenzymatic process in which proteins react with reducing sugar molecules and thereby impair the function and change the characteristics of the proteins. Glycation is involved in diabetes and aging where the accumulation of glycation products causes side effects. In this study, we statistically investigate the glycation of epsilon amino groups of lysines and also train a sequence-based predictor. The statistical analysis suggests that acidic amino acids, mainly glutamate, and lysine residues catalyze the glycation of nearby lysines. The catalytic acidic amino acids are found mainly C-terminally from the glycation site, whereas the basic lysine residues are found mainly N-terminally. The predictor was made by combining 60 artificial neural networks in a balloting procedure. The cross-validated Matthews correlation coefficient for the predictor is 0.58, which is quite impressive given the relatively small amount of experimental data available. The method is made available at www.cbs.dtu.dk/services/NetGlycate-1.0.
糖基化是一种非酶促过程，在此过程中蛋白质与还原糖分子发生反应，从而损害蛋白质的功能并改变其特性。糖基化与糖尿病和衰老有关，在这些情况中糖基化产物的积累会产生副作用。在本研究中，我们对赖氨酸的ε氨基的糖基化进行了统计调查，并训练了一个基于序列的预测器。统计分析表明，酸性氨基酸（主要是谷氨酸）和赖氨酸残基催化附近赖氨酸的糖基化。催化性酸性氨基酸主要在糖基化位点的C端被发现，而碱性赖氨酸残基主要在N端被发现。该预测器是通过在投票程序中组合60个人工神经网络构建而成。该预测器的交叉验证马修斯相关系数为0.58，鉴于可用的实验数据量相对较少，这一结果相当可观。该方法可在www.cbs.dtu.dk/services/NetGlycate-1.0获取。
