Cluster Analysis Based On Single Cell Development Data Of C.elegans

In order to study the embryonic development of fertilized eggs,biologists used 4D confocal imaging technology to quantitatively observe the gene expression dynamic behavior of Caenorhabditis elegans at the single cell level.This kind of data has the following three characteristics:1)spatiotemporal heterogeneity;2)high noise;3)tree structure.This paper aims to explore the gene expression patterns in the development process by developing statistical clustering algorithm.First of all,we try to cluster all kinds of genes using gene expression data.According to some problems in the real data,we deal with the data,abstract the process of cell division and differentiation into a binary tree model,and construct the similarity measure of binary tree.This paper shows the process of method selection and construction,and applies the model to synthetic data and real data.Its performance in synthetic data is better than other methods,and its application results in actual data are consistent with our cognition.