学院总览

我院Kuan Yoow Chan课题组交叉研究成果在Bioinformatics上发表

时间:2022-10-10 阅读量:1 来源:浙江大学爱丁堡大学联合学院

        近日,浙江大学爱丁堡大学联合学院(ZJE)Kuan Yoow Chan课题组在Bioinformatics上发表题为pcnaDeep: A Fast and Robust Single-Cell Tracking Method Using Deep-Learning Mediated Cell Cycle Profiling的应用程序简介(Application Note),介绍了课题组开发的基于深度学习和细胞周期识别的细胞追踪程序pcnaDeep。ZJE生物信息学18级本科生归逸凡(现剑桥在读博士)和ZJE双学位博士生谢双双为本文共同一作。本研究在ZJE Kuan Yoow Chan老师和浙江大学伊利诺伊大学厄巴纳香槟校区联合学院(ZJUI)王高昂老师的共同指导下开展,将计算机算法技术应用于生物医学研究领域,是浙江大学国际校区多学科交叉学术氛围下的可喜成果。


      近年来,单细胞水平的定量分析引领了关于细胞命运分子机制研究方法的革新。然而,在利用显微成像得到的时间序列数据开展活体单细胞分析方面,研究者仍受限于细胞追踪方法的准确度。尤其对于较难追踪的细胞分裂事件而言,误差较大的追踪结果意味着需要大量人力介入开展纠正工作。


       在本研究中,研究人员利用近年来应用较广的细胞周期标志物——荧光标记的PCNA蛋白,使用深度学习模型Mask R-CNN学习PCNA荧光蛋白在不同细胞周期特有的空间分布特征,从而实现对细胞周期的识别。在特征设置上,将荧光与明场图像结合有助于提高处于分裂期细胞的识别准确度。随后,通过将细胞周期信息与传统目标追踪算法结合,研究人员得以准确链接细胞分裂前后细胞的轨迹。这一方法能够使研究人员迅速准确地得到显微成像中的单细胞轨迹,有助于研究特定蛋白在细胞中的时空动态变化。

图片 1.png

Quantitative single-cell analysis has revolutionised our approach to study the molecular events that lead to cell fate decisions. However, a common challenge when performing single-cell analysis is the limited computational tools available to track single-cells over multiple cell generations robustly. This is because when cells undergo cell division, they form daughter cells that can confound the algorithms used to track the cells, leading to high false mother-daughter assignments. This means extensive manual corrections are often required, making it a time-consuming practice.


In the recent publication in Bioinformatics, Kuan Yoow Chan's lab developed an application that uses deep learning to classify cells based on the pattern of the well-established fluorescently tagged PCNA protein. By using the PCNA pattern to identify the cell cycle state of cells, we were able to assign mother-daughter relationships rapidly and accurately. This novel approach has high accuracy rates and relatively low computational costs, allowing us to generate high-quality single-cell lineages from large long-term imaging datasets.

    原文链接:

https://academic.oup.com/bioinformatics/advance-article/doi/10.1093/bioinformatics/btac602/6680181


课题组介绍

About lab

Kuan Yoow Chan课题组利用遗传学、细胞生物学和生化分析,旨在研究中心体信号的变化如何导致癌症中细胞周期的反常。Using genetic, cell biology and biochemical assays, Kuan Yoow Chan' Labaim to study how the changes in centrosomal signalling lead to the deregulation of the cell cycle in cancer.


Email:kychan@intl.zju.edu.cn

实验室网址:https://person.zju.edu.cn/H118031