基于计算方法的抗菌肽预测

发布时间：2018-05-28 15:07

本文选题：抗菌肽预测 + 计算方法　；参考：《计算机学报》2017年12期

【摘要】：抗菌肽是由生物体免疫系统所产生的能抵抗微生物感染的一种小分子多肽,因其具有高效低毒的广谱抗菌活性且几乎无耐药性问题,被看做是抗生素的最佳替代品,对解决抗生素滥用问题具有重要的意义.抗菌肽预测是生物信息学的一个重要研究内容,对抗菌肽及其抗菌功能进行预测能有效帮助了解抗菌肽的作用机理,为抗菌肽药物的设计和改造提供理论依据.基于计算方法的抗菌肽预测是采用数学理论、计算机技术和生物信息学方法,通过对抗菌肽数据的分析来挖掘出抗菌肽的生物特征和抗菌活性之间的关联,从而自动地对抗菌肽的类别做出推断.由于不依赖于生物实验,而是依靠有效的算法和计算机的高速计算能力来完成预测工作,计算方法具有高效快捷、成本低廉等特点,且具有良好的可操作性和批量处理能力,非常适合大规模预测任务,因此已经引起了国内外学者越来越多的关注.文中对国内外的相关研究成果进行了阐述和总结,包括抗菌肽生物信息数据库、主流的预测方法和预测方法的性能检验等.抗菌肽数据库是专门针对抗菌肽建立的数据库,收录了大量的抗菌肽数据,使用者不仅可以从中提取所需要的信息,还可以使用数据库所提供的各类在线工具对数据进行处理.文中对常见的一些抗菌肽数据库进行了介绍,给出相关数据库的数据收录情况、功能特点和网址链接等,以方便读者查询使用.接着文中介绍了目前主要使用的抗菌肽预测方法,包括基于经验分析的预测方法和基于机器学习的预测方法,前者是根据已知的经验规则或者模式对某类抗菌肽的一些生化属性和抗菌活性之间的关联进行统计或建模来对该类抗菌肽进行识别,而后者则是利用机器学习技术,通过对抗菌肽的已知数据信息进行学习,建立合理的预测算法从中找出抗菌肽的特点和规律,并将其推广到未知多肽数据来进行预测.随后文中又给出了预测方法的评估方法和评价指标,这些性能检验结果既是评估一个方法预测性能好坏的标准,又是与其他方法进行比较的依据.最后,文中对抗菌肽预测的发展进行了思考和讨论,并展望了未来的研究方向.
[Abstract]:Antimicrobial peptides are small molecular peptides produced by the immune system of organisms that can resist microbial infection. Because of their high efficiency and low toxicity, antimicrobial peptides are considered as the best substitute for antibiotics because of their wide spectrum antibacterial activity and almost no drug resistance. It is of great significance to solve the problem of antibiotic abuse. The prediction of antimicrobial peptides is an important research content in bioinformatics. The prediction of antimicrobial peptides and its antibacterial function can effectively help to understand the mechanism of antimicrobial peptides and provide theoretical basis for the design and modification of antimicrobial peptides. The prediction of antimicrobial peptides based on computational methods is based on mathematical theory, computer technology and bioinformatics methods. By analyzing the data of antimicrobial peptides, we can find out the relationship between the biological characteristics and antibacterial activities of antimicrobial peptides. This automatically inferred the types of antimicrobial peptides. Because it does not depend on biological experiment, but relies on effective algorithm and high speed computing ability of computer to complete the prediction work, the calculation method has the characteristics of high efficiency, high speed, low cost, and has good maneuverability and batch processing ability. It is very suitable for large-scale prediction task, so it has attracted more and more attention from scholars at home and abroad. In this paper, the related research results at home and abroad are described and summarized, including the biological information database of antimicrobial peptides, the main prediction methods and the performance test of prediction methods. Antimicrobial peptide database is a database specially established for antimicrobial peptides, which contains a large number of antimicrobial peptide data. Users can not only extract the needed information from it, but also process the data by using all kinds of online tools provided by the database. In this paper, some common antimicrobial peptide databases are introduced, and the data collection, function features and URL links of the related databases are given, so as to facilitate the readers to inquire and use. Then, the paper introduces the main prediction methods of antimicrobial peptides, including empirical analysis based prediction and machine learning based prediction methods. The former uses known empirical rules or patterns to identify the relationship between some biochemical properties and antibacterial activities of a class of antimicrobial peptides, while the latter uses machine learning technology. By learning the known data information of antimicrobial peptides, a reasonable prediction algorithm is established to find out the characteristics and rules of antimicrobial peptides, and to extend them to unknown peptide data to predict. Then the evaluation methods and evaluation indexes of the prediction methods are given. These performance test results are not only the criteria for evaluating the performance of one method, but also the basis for comparison with other methods. Finally, the development of antimicrobial peptide prediction is discussed, and the future research direction is prospected.
【作者单位】：大连理工大学控制科学与工程学院;
【基金】：国家自然科学基金(61502074) 中国博士后科学基金资助项目(2016M591430) 大连理工大学基本科研业务费科研项目(DUT17RC(4)09)资助~~
【分类号】：Q811.4;TP311.13

【相似文献】