《生命科学》 2019, 31(4): 357-363

数据与计算驱动的蛋白质元件预测和设计

刘海燕*，陈　泉，龙朋朋，黄　斌，许　洋，陈耀晞

(中国科学技术大学生命科学学院，大数据学院，合肥 230027)

摘　要：摘　要：为实现特定合成生物系统，需要使用恰当的蛋白质元件，即具有所需的特异性分子识别、酶催化活性等功能的天然蛋白质或工程改造蛋白质。以转录因子所识别的DNA 序列的预测以及蛋白质- 小分子特异性结合口袋的预测和设计为例，介绍计算方法在蛋白质功能预测和设计中的作用。强调了不同类型计算工具的整合以及它们与生物背景知识整合、计算方法通用性和准确性之间的平衡；讨论了有待解决的问题、计算的潜力和新方法的发展需求。

Data and computation-driven prediction and design of protein components

LIU Hai-Yan*， CHEN Quan， LONG Peng-Peng， HUANG Bin， XU Yang， CHEN Yao-Xi

(School of Life Sciences， School of Big Data， University of Science and Technology of China， Hefei 230026， China)

Abstract: Abstract: To implement a particular synthetic biology system， one needs appropriate protein components， namely， natural or engineered proteins that possess required specific molecular interactions， catalytic activities， and so on. Here we consider as examples the prediction of DNA sequences recognized by transcriptional factors and the prediction and design of protein-small molecule-binding pockets， to illustrate the roles of computation in the prediction and the design of protein functions. We emphasize the integration of different types of computational tools and of the tools with background biological knowledge， and the balance between the generalizability and the accuracy of computational methods. We also discuss important unsolved problems， the potential of computation， and the needs for new methods.