人工设计蛋白质是深入认识蛋白质序列结构规律的新途径，也能为基于合理设计的蛋白质工程提供工具。蛋白质设计可分为两个层次：以给定主链折叠为目标设计氨基酸序列和从头设计有高”可设计性“的主链折叠。这里“可设计性”是指能够稳定折叠成给定主链结构的氨基酸序列确实存在。对主链折叠“可设计性”的实验验证依赖于高成功率的氨基酸序列设计方法。我们在前期研究中建立了氨基酸序列设计方法，并进行了实验验证。本项目将在对全自动序列设计进一步完善的基础上，采用链生长抽样、多尺度模型等技术，发展能设计蛋白质主链折叠的模拟技术；以天然蛋白质结构数据为依据，基于贝叶斯理论建立”可设计性“的评价函数。最后，以人工构建的主链折叠为目标结构进行氨基酸序列设计，获得主链折叠和氨基酸序列均为人工设计的新蛋白。

Protein design is a new route to deepening our understanding of rules governing protein sequence-structure relations. It also provides tools for protein engineering based on rational design. The task of protein design can be decomposed into two levels, including the design of amino acid sequences for a given target backbone structure or backbone fold, and the design of backbone folds that are associated with high designablility. The designability of a backbone fold refers to that amino acid sequences that stably fold into this backbone structure actually exist. To experimentally verify the designability of a backbone fold relies on amino acid sequence design methods that have high success rate. We have developed and experimentally verified a method for sequence design. In current project, based on further improvement of this method, we will develop simulation techniques including chain growth and multi-scale modeling methods that can design backbone folds. We will also develop scoring functions for designability based on known protein structure data and the Bayesian theory. Finally, amino acid sequences will be designed for artificially constructed backbone folds. Such designed proteins will be experimentally verified to fold into respective designed backbone structures.