报告题目：Deep learning prediction boosts phosphoproteomics-based discoveries through improved phosphopeptide identification
报 告 人：伊心培
工作单位：上海交通大学
报告时间：2022-12-12 19：00-21:00
腾讯会议ID： 723-823-545
报告摘要：
Shotgun phosphoproteomics has been widely used for phosphorylation study. The phosphorylation identification from phosphoproteome datasets relies largely on the database search strategy and phosphosite localization. Search space inflation and site-specific peak missing leads to a low sensitivity of current phosphorylated peptide identification. In order to improve the sensitivity of phosphorylated peptide identification, we proposed DeepRescore2 which is a novel post-processing tool by combining deep learning derived predictions, retention time and spectrum similarity, to facilitate the phosphosite localization and rescore peptide spectrum matches. The method is benchmarked on phosphopeptide synthetic dataset and is applied to process real label free and TMT phosphoproteome datasets. It is shown that DeepRescore2 supplements the probability-based phosphosite localization and improves the sensitivity of phosphorylated peptide identification on different datasets combined with different search engines.
报告人简介：
伊心培，博士毕业于中国科学院数学与系统科学研究院，导师巩馥洲研究员。现任上海交通大学生命科学技术学院助理研究员，研究方向为基于癌症蛋白基因组数据的生物信息学，计算生物和生物统计方法开发和分析。在 Iscience,Journal of Proteome 和 BMC Bioinformatics等杂志发表多篇论文。
