Reveal heterogeneous motion states in single nanoparticle trajectory using its own history

作     者：Hansen Zhao Feng Ge Sichun Zhang Xinrong Zhang Yan He Hansen Zhao;Feng Ge;Sichun Zhang;Xinrong Zhang;Yan He

作者机构：Department of ChemistryTsinghua UniversityBeijing100084China 

出 版 物：《Science China Chemistry》 (中国科学（化学英文版）)

年 卷 期：2021年第64卷第2期

页      面：302-312页

核心收录：

学科分类：07[理学] 070205[理学-凝聚态物理] 08[工学] 080501[工学-材料物理与化学] 0805[工学-材料科学与工程（可授工学、理学学位）] 0702[理学-物理学] 

基　　金：the National Natural Science Foundation of China(21425519 21221003) 

主　　题：single particle tracking nanoparticle machine learning 

摘      要：Single particle tracking(SPT)has long been utilized for investigation of complex system dynamics such as nanoparticle-cell interaction,however,the analysis of individual particle motions is always a difficult *** methods treat each data point or fragment on the recorded trajectory as an isolatedatomand determine their relationship based on externally predefined models or physical states,which inevitably lead to oversimplification of the associated spatiotemporal ***,inspired by the historical analysis in social science,we propose a modeless preprocessing framework for SPT analysis based on thehistoryof the *** new strategy consists of 3 steps:(1)assign ahistoryto each data point and construct successive overlapped historical vectors;(2)perform unsupervised clustering in the vector space to find their relative differences;(3)project differences back to the trajectory by coloring each point accordingly for *** a result,the inner heterogeneity of the particle motion self-emerges as a colored trajectory,exhibiting a global picture of the local state transitions and providing valuable information for further model-based *** the complexity issues at various spatiotemporal scales have attracted increasing attention,and individual objects such as single molecules,cells,vehicles and even stars in the universe could all be treated assingle particles,this presuppositionless data preprocessing approach could help the investigations of many complex systems in fundamental research.