The Study Of Automatic Classification Of Protein Folding Type

Studying the protein folding is an important topic in the field of life science.The classification,recognition,structure and function prediction of folding types are sequential and the classification of protein folding types is the basis for the study of folding rules.Protein folding type reflects the core of protein structural topology model,it is based on protein secondary structure unit of a form of description,it covers protein spatial structure composed of three aspects,including the secondary structure units,the relationship between the relative arrangement of secondary structure unit and the peptide chain to the protein polypeptide chain.The existence of thousands of types of protein folding system classification and recognition,exploring the rule of formation of protein folding type,will help to reveal the protein folding pattern,and provide accurate protein tertiary structure and function prediction.The research content includes the following aspects:1.The protein folding type template design and template database constructionSelection of template is the basis of protein folding type classification,the validity of template directly affect the effect of classification.Selection of AstralSCOPe 2.05 database similarity in is less than 40%,and the resolution is higher than 0.25 nm All alpha proteins(alpha),All beta proteins(beta),alpha and beta proteins(alpha/beta),alpha and beta proteins(alpha+beta)989 kinds of subordinate to the four types of protein folding type as the research object,based on protein structure the results between the sample and data analysis,the folding type family template design method was built.Template and the family as the unit by system clustering to build preliminary protein folding type template,based on system clustering nodes in the graph corresponding to the calculation analysis and inspection of initial template,and template selection standard of experience of any protein folding type was proposed,the design method of protein folding type template was established.Respectively using the above template design method,a template database which contains 3941 family template and a folding type template database with a total of1617 fold type were built.2.The protein folding type classification method based on template databaseIn this paper,using the structure method of TM-align with max TM-score points,the type classification method based on template database of protein folding was established.The basic idea is: By comparing any protein sample under test with all templates in template data with TM-align,and calculate the TM-score value,the folding type of template with max TM-score is the folding type of the sample to be measured.Classification results using the sensitivity,specificity,Matthew correlation coefficient of three indicators to evaluate it.To test and verify the rationality of template design and classification method and universality,respectively using family template database with folding type template database for protein folding type classification of self consistent inspection and independence.The test shows that based on family template database self consistent test results the sensitivity,specificity and MCC was 95.00%,99.99% and 0.94 respectively,and the average based on folding type template database self consistent test results the sensitivity,specificity and MCC's average of 93.71%,99.97% and 0.91 respectively.Two types of templates for the same data set the classification of the test results,the classification results of the former slightly higher the latter.That a family template and folding type template design is reasonable,the template reflects the basic feature of folding type;Template for a total of 3941 of the former,the latter is only 1617,which the template is only two 5 of the former,classification speed of the latter is much better than the former,classification accuracy family template slightly better than folding type template.The independence test shows that the family type template database template database and folding effect on expansion of the classification of the sample is a bit poor in self consistent test results,but in the family independence test template and the classification of the folding type template effect is generally higher than 90%,shows the template database and classification method can be used to extend the protein fold type classification samples,confirming the universality of the template design and classification method is effective.In this paper,989 types of protein folding were studied systematically,the design method of protein family and protein folding type template was established,the construction of protein family and protein folding template database was completed;Based on template established protein folding type classification method,automatic classification of protein folding types is realized.