上海科技大学知识管理系统

本发明提供了一种用于深度学习判别模型训练的功能多肽数据库，属于数据库技术领域。包括：数据收集模块：用于通过检索互联网检索相关的多肽数据集信息；数据汇交模块：用于将积累的已标注功能多肽数据集进行分类，总共分为七类，并对每一个功能肽库，排查重复序列(库频>1)，并只保留只含有20种天然氨基酸残基的多肽序列；数据存储模块；数据应用模块。该发明在对数据库的建立过程中大量使用基于python编程和爬虫技术，以此获取公有肽数据源上目标功能肽的序列信息；本数据库以csv格式存储每条多肽的序列及其生物功能信息，方便后期用于基于NLP自然语言的深度学习算法的运用，并且可以有效地帮助研发人员快速对众多候选多肽进行优先级。

The invention provides a functional polypeptide database for deep learning discrimination model training, and belongs to the technical field of databases. Includes : data collecting module for searching related polypeptide data set information through searching internet; Data convergence module : used to classify the accumulated labeled functional peptide data set, divided into seven categories in total, and for each functional peptide library, check the repetitive sequence (library frequency > 1), and only retain only 20 kinds of natural amino acid the peptide sequence of residues; A data storage module; Data application module. In the process of building the database, python-based programming and crawler technology are widely used to obtain the sequence information of the target functional peptide on the public peptide data source; This database stores the sequence of each polypeptide and its biological function information in csv format, which is convenient for later use in deep learning algorithms based on NLP natural language, and can effectively help research and development personnel to quickly prioritize numerous candidate polypeptides.