Abstract

Info

Author(s):: LIU Zinan; LI Heshan; SONG Xiaoyu; School of Life Science and Technology, Harbin Institute of Technology, Harbin 150080, China

Keywords:: protein structure prediction deep learning homology modeling ab initio prediction review

Abstract:: The prediction of protein structure has provided important underpinnings for understanding the biological functions of protein at molecular level. Herein the development of protein structure predication is reviewed from the aspects of homology modeling, ab initio prediction and deep learning. Because of the increasing number of protein templates with known structure, the improved performance of sequence alignment algorithm for information extraction and the application of fragment assembly technology, the ability of homology modeling for predicting protein structure is greatly increased. The prediction of long-distance amino-acid contact using ab initio method is more and more precise for the application of regional partition and fragment segmentation and the strategy of parallel computing in ab initio method. Deep learning combined with the above-mentioned classical technologies can promote the accuracy and speed of protein structure prediction, but there is still a great challenge when the protein has no homologous proteins.