软件导刊2025，Vol.24Issue(8)：11-16,6.DOI:10.11907/rjdk.241459
基于深度学习的蛋白质从头测序方法研究进展
Research Progress on Protein De Novo Sequencing Methods Based on Deep Learning
摘要
Abstract
One of the core tasks of proteomics is to reveal the amino acid sequence of proteins.It has become the dominant means of large-scale protein research.The protein de novo sequencing method can directly analyzes tandem mass spectrometry data without relying on protein database information.Due to its ability to recognize new proteins and provide post-translational modification information,it is highly favored.The fusion of DeepNovo series,Denovo GCN,GraphNovo and other models provides diverse and innovative methods for de novo peptide se-quencing of proteins.Combining the powerful capabilities of deep learning and the needs of proteomics,it opens up new ways to improve the accuracy and efficiency of protein identification,providing more accurate and reliable tools for bioinformatics research.Although de novo se-quencing methods for proteins have unique advantages,their identification accuracy is still affected to some extent due to the complexity of tan-dem mass spectrometry data and the limitations of the method itself.After summarizing the development process of protein de novo sequencing methods in depth,explore the advantages and limitations of emerging methods based on deep learning.These methods fully utilize the advan-tages of deep learning in analyzing tandem mass spectrometry information.By continuously optimize algorithm design and deep learning mod-els,in order to achieve the goal of improving the accuracy of protein identification.关键词
蛋白质从头测序/蛋白质组学/串联质谱技术/深度学习Key words
protein de novo sequencing/proteomics/tandem mass spectrometry technology/deep learning分类
信息技术与安全科学引用本文复制引用
荆昊,贾晓田,柳楠..基于深度学习的蛋白质从头测序方法研究进展[J].软件导刊,2025,24(8):11-16,6.基金项目
国家自然科学基金青年科学基金项目(61902221) （61902221）
