Abstract
Advances in microtechnologies are making it possible for high-throughput control and reporting of gene expression in live cells, in real-time. We explore relevant statistical challenges to modeling and inference in real-time gene expression data from single-shock experiments, with special attention on potential confounding between treatment and cell cycle variation. We propose a semi-wavelet non-linear dynamic regression model to infer modulation in gene expression due to treatment shocks in the presence of cell cycle variation. A case study is performed with public data. Results are compared ignoring cell cycle. Estimation and inference are performed by a Bayesian approach.