Gene regulatory network inference using multivariate information measures Seminar

Event details

Single cell transcriptomic data are increasingly used to study the dynamics of gene expression during cell development. We have developed an algorithm, based on multivariate information measures, to identify (co-)regulatory relationships between genes from these data. Information measures allow us to discern non-linear statistical relationships among random variables that are typically missed by conventional correlation measures. We compare our method to existing information theoretical-based network inference algorithms and discuss factors that influence their performance. Using experimental data from mouse embryonic stem cells, we illustrate how we can use our method to identify changes in gene interactions over time as cells undergo differentiation.