Estimation of cell state dynamics by deep generative model

Most omics observations, including single-cell transcriptome observations, are invasive and provide only snapshot observations. We are developing a methodology to recover the stochastic dynamics of the cellular state by integrating splicing mathematical models with deep generative models. By doing so, we aim to elucidate the molecular mechanisms of cell state transitions, such as the search for molecular mechanisms at work during the generation of highly malignant tumor cells, and to propose molecular interventions to prevent the progression of cancer.