A method for seamless alignment of multi-domain protein templates

Abstract

The invention discloses a polydomain protein template seamless comparison method. Starting from a single-domain protein three-dimensional structure, the method comprises the steps that firstly, blank sequences with the allowable length are added to the two ends of a template; then, a protein template is adopted for searching a tool TM-align for the best comparison position of a first single-domain protein on a template and searching the portion, behind the last comparison position, of the template for the best comparison position of a second single-domain protein, if a gap between the two comparison portions is larger than the maximum allowable value, a seamless sliding window method is adopted for searching for the best comparison position of the second single-domain protein, and after the best comparison position of the second single-domain protein is determined, an identical method continues to be adopted for searching for the best comparison positions of other single-domain proteins; finally, the best comparison position of each single-domain protein is adopted as the best comparison between the current search protein and the target. The polydomain protein template seamless comparison method is high in comparison precision and efficiency.

Description

technical field [0001] The invention relates to the fields of biological informatics, intelligent optimization and computer application, and in particular to a method for seamless comparison of multi-domain protein templates. Background technique [0002] A protein may contain one domain or multiple domains. A protein containing only one domain is called a single-domain protein, and a protein containing two or more domains is called a multi-domain protein. The three-dimensional structure prediction of multi-domain proteins is an important step in understanding the function of full-length proteins. Since multi-domain proteins are generally larger in scale than single-domain proteins, it increases the difficulty of predicting the structure of multi-domain proteins. At present, the most commonly used multi-domain protein structure prediction methods start from the single-domain structure in multi-domain proteins, and predict the structure of multi-domain proteins by sampling th...

AI Technical Summary

Problems solved by technology

If the template comparison is inaccurate, even if an excellent template is found, it is impossible to predict a structure with higher accuracy

Since multi-domain proteins are composed of multiple single-domain proteins, and the length of the connection between each single-domain protein corresponding to the connection on the template cannot be known in advance, it is not possible to use the comparison method of single-domain proteins for comparison

[0004] At present, the commonly used multi-domain protein alignment method is based on the threading alignment method. However, the threading alignment starts from the entire sequence of the protein. During the alignment process, the length of the connection between the domains cannot be adjusted to achieve Optimum, so that templates with similar structures to the query protein but with different connection lengths cannot be aligned to the optimal position, thus affecting the alignment accuracy

[0005] Therefore, existing multi-domain protein template alignment methods are deficient in alignment accuracy and need to be improved

Images