文献类型：期刊文献

英文题名：Analysis and classification of nanopore data based on feature-level multi-modality

作者：Fu, Xixin[1]; Wan, Yongjing[1]; Li, Xinyi[2]; Ying, Yilun[2]; Long, Yitao[2]

机构：[1] East China University of Science and Technology, College of Information Science and Engineering, Shanghai, China; [2] Nanjing University, School of Chemistry and Chemical Engineering, Nanjing, China

年份：2020

起止页码：692

外文期刊名：Proceedings - 2020 13th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics, CISP-BMEI 2020

收录：EI(收录号：20210409816832)

语种：英文

外文关键词：DNA sequences - Gene encoding - Nanopores - Spectrum analysis - Classification (of information) - Time domain analysis

摘要：As a powerful single-molecule analysis technology, nanopore analysis is used in DNA sequencing and biomolecular recognition, providing a powerful analytical method for biochemistry. Traditional signal recognition methods of nanopore is based on the statistical characteristics of electrical signals, which can only distinguish signals with obvious differences in time domain characteristics. The introduction of spectrum enables effective classification of similar signals from nanopores, which is of great significance to its further application. In this paper, we design the ensemble empirical mode decomposition, variational mode decomposition, inherent time scale decomposition and Hilbert transform to extract multi-spectral features of nanopore electrical signals. Then, the signal feature spectrum can be classified and identified through ResNet network. The results prove that the feature-level multi-modality fusion method can improve the classification and recognition effect of similar electrical signal spectrum features, which is significantly promote the sensitivity of nanopore technology in single-molecule analysis. ? 2020 IEEE.