摘要
传统抗微生物肽识别分析主要通过实验手段进行,效率低,耗费较多人力物力。最新的抗微生物肽识别方法是将计算机技术和生物信息学相结合,通过机器学习方法进行大数据挖掘分析,从大量的多肽序列数据里面预测抗微生物肽,从而加快抗微生物肽的识别。收集并分类整理了近10年来计算机辅助抗微生物肽识别的研究文献,从中梳理出抗微生物肽的主要数据资源、抗微生物肽识别的特征工程、抗微生物肽的机器学习预测算法和抗微生物肽的回归分析方法。同时,进一步对机器学习算法的模型性能评估方法进行综述,总结其中存在的不足并展望了未来的发展方向。
The traditional methods for the identification of antimicrobial peptides are experimental means,which is inefficient and consumes a lot of manpower and material resources.The latest ways to identify antimicrobial peptides combine computer technology,bioinformatics,and machine learning methods together.Based on big data mining and analysis,antimicrobial peptides can be predicted from a large amount of peptide sequence data.The identification of antimicrobial peptides thereby could be accelerated.This paper classifies and analyzes the main literatures of the computer-aided antimicrobial peptide recognition in the recent 10 years,sorts out the main antimicrobial peptides data resources,the characteristic engineering of antimicrobial peptide recognitions,the machine learning prediction algorithms of antimicrobial peptides,the regression analysis methods of antimicrobial peptides.In the meanwhile,this paper reviews the model performance evaluation methods of machine learning algorithms,summarizes the existing shortcomings,and prospects the future development directions.
作者
刘明友
刘红美
张招方
朱映雪
黄健
LIU Mingyou;LIU Hongmei;ZHANG Zhaofang;ZHU Yingxue;HUANG Jian(School of Biology and Engineering,Guizhou Medical University Guiyang 550025;School of Life Science and Technology,University of Electronic Science and Technology of China Chengdu 610041;Taiheyun Engineering Consulting Co.,Ltd.Guiyang550081)
出处
《电子科技大学学报》
EI
CAS
CSCD
北大核心
2022年第6期830-840,共11页
Journal of University of Electronic Science and Technology of China
基金
国家自然科学基金(62071099)
贵州省卫生健康委科学技术基金(gzwkj2023-590)。
