Recognition of Protein Native Folds Using a Continuous Potential Function

e report a simple continuous potential function that can recognize proteinchains' native folds from tens of thousands of alternative ones. Empirical parame-ters for this potential function were obtained by a neural network learning oversamples generated from PDB structural data.Hydrophobic interactions were foundto be mainly responsible for stabilization of the protein' s native fold. The inter-chain interaction was found indispensable in stabilizing some protein chains' nativeconformation.

蛋白质天然构象预测的研究进展

继人类基因组计划完成后,如何弄清相应的基因序列的功能及其之间的关系,即比较彻底地解开生命的奥秘成为后基因组计划的主要任务。基因的功能最终主要通过蛋白质来完成,因此,如何根据已知的氨基酸序列来预测相应的蛋白质天然构象成为后基因组计划中最重要的组成部分之一,这为比较彻底界定其功能奠定分子生物学基础。本文从以下几个方面进行了详细地介绍:(1)组成蛋白质天然构象预测方法的基本元素即序列比对方法、构象空间搜索方法及有关的蛋白质数据库;(2)结构预测方法的评估与分类;(3)描述预测性能较好及有比较发展前景的预测方法;(4)天然构象预测的前景等。