基于深度学习的蛋白质亚细胞定位算法概述  

Protein Subcellular Localization Algorithm Based on Deep Learning

作　　者：谭博 Tan Bo(Guangdong University of Technology,Guangzhou Guangdong 510006)

机构地区：[1]广东工业大学,广东广州510006

出　　处：《科学与信息化》2020年第3期137-138,共2页Technology and Information

摘　　要：蛋白质亚细胞定位是分子细胞生物学和蛋白质组学的一个重要课题,有助于了解蛋白质的性能和功能。荧光显微镜法是亚细胞定位最常用的方法和手段,常见方法有专家观测荧光图像标注蛋白质、利用形态学特征提取自动识别以及神经网络提取特征等,但多标记荧光蛋白质细胞图像的多标签分类任务其准确度还远低于人工观测。subcellular localization of proteins is an important subject in molecular cell biology and proteomics,which is helpful to understand the properties and functions of proteins.Fluorescence microscopy is the most commonly used method and means of subcellular localization.Common methods include expert observation of fluorescent image to label protein,automatic recognition by morphological feature extraction and feature extraction by neural network,etc.However,the accuracy of multi label classification task of multi label fluorescent protein cell image is far lower than that of manual observation.

关 键 词：蛋白质亚细胞定位 神经网络 多标签分类 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]