A non-negative multilinear block tensor decomposition approach to flow cytometry data analysis

The paper presents a novel approach to the processing of flow cytometry data sequences. It consists in decomposing a sequence of multidimensional probability density functions by using multilinear block tensor decomposition approach [1]. The identifiability of the model is also addressed as well as the data processing. To illustrate the effectiveness of the approach, a study of the T47D cell line mitochondrial membrane potential as a function of the CCCP¹ decoupling agent concentration is performed. The cell sorting capacity of the method is significantly improved as compared to classical clustering methods.