Research On The Prediction Of Protein Structure Based On Homologous Modeling

Obtaining structure and function of proteins is one of the main purposes of research in biology, but determining three-dimensional structure of protein by means of X-rays crystallography and nuclear magnetic resonance spectroscopy, and research in protein structure by the way of biochemistry, already can't adapt the need of the rapid increase of protein sequence. In bioinformatics, the structure prediction of new sequence can be used to establish the relation between sequence and corresponding to the structure as well as function by making use of the structure-known model of the protein family, so that we can predict the structure and function of protein in less cost and shorter time.This paper mainly concentrates on an aspect that we can predict three-dimensional structure of a protein which is given originally through homologous modeling. This process is composed of such sections as recognition of target template, alignment between query sequence and target sequence, modeling, evaluation of structure rationality. And among those sections, recognition of target template is completed by BLAST(Basic Local Alignment Search Tool),modeling is accomplished by modeling software MODELLER, evaluation of structure rationality is achieved by software PROCHECK. Additionally, in the whole process of protein structure prediction, alignment between query sequence and target sequence affects modeling greatly ultimately, so it plays an very important role for the whole process of protein structure prediction to improve precision of alignment between query sequence and target sequence.The paper pays great attention to how to improve the precision of alignment between query sequence and target sequence. At present, in various way of structure prediction of new sequence, there lie such problems: in the process of modeling a new sequence, the alignment between target sequence and model sequence we get is not very accurate, so it has a negative effect on the protein structure predicted ultimately. According to the above, this paper mainly does research on three aspects: using profile-profile method in the process of getting alignment between target sequence and model sequence, data smooth we use in...