Deng, Dianliang; Fang, Hong-Bin; Razeghi Jahromi, Kian; Song, Jiuzhou; Tan, Ming Detection of threshold points for gene expressions under multiple biological conditions. (English) Zbl 1388.62137 Stat. Interface 10, No. 4, 643-655 (2017). Summary: Temporal gene expression data is of importance in the classifications of gene functions and have been extensively used in biomedical studies, such as cancer diagnostics. However, since temporal gene expressions vary over time, after the initial time periods, many genes exhibit some kind of stability, which means that gene expressions keep constant or fluctuate slightly after those time points. Thereby, this threshold point is a key in the study of behaviours of gene expressions, which can be used to decide the measuring time period and to distinguish the gene expressions. In this paper, three methods are presented to detect the threshold points for the gene expressions. In particular, the first-order and second-order change rates are used to construct the test statistics for detecting the threshold points. The simulation study shows that the proposed methods have a good performance for the detection of threshold points. A real dataset with 21 genes in P. aeruginosa expressed in 24 biological conditions is used to illustrate the proposed methodology. MSC: 62G20 Asymptotic properties of nonparametric inference 62H15 Hypothesis testing in multivariate analysis 62P10 Applications of statistics to biology and medical sciences; meta analysis Keywords:relative change rate; temporal gene expression; empirical distribution; high dimensional data PDFBibTeX XMLCite \textit{D. Deng} et al., Stat. Interface 10, No. 4, 643--655 (2017; Zbl 1388.62137) Full Text: DOI