Tobias Wängberg: The Statistical Modelling of Gene Transcriptional Burst Kinetics
Time: Wed 2022-02-23 15.15 - 16.00
Location: Kräftriket, House 6, Room 306
Participating: Tobias Wängberg (Stockholm University)
Technological developments during the previous decades allow collection of data on the abundance of gene expressions at single cell resolution. A common assumption when analysing such data is that the expression counts follow a negative binomial (NB) distribution. Although the NB distribution is able to model the data well in many cases, any justification of this assumption, other than the observed overdispersion, is often lacking. Moreover, biological interpretation of the parameters are nontrivial. In this talk I will introduce a Markovian model that justify and generalize the NB distribution as a limit of the Poisson-Beta mixture stationary distribution, based on the assumption that mRNA synthesis occur in short bursts. I will also discuss several biological and statistical insights that can be gained from the model, such as a clear interpretation of model parameters that connects back to the underlying biological mechanism. Lastly I will outline how the model can help us identify different cell types based on the corresponding gene expression profiles.