Professeur Zoran Nikoloski

University of Potsdam and Max-Planck Institute for Molecular Plant Physiology

Computational efforts in the last decade have led to two different types of approaches for inference of gene regulatory networks (GRNs) given large-scale gene expression data, namely unsupervised and supervised. I will present an approach based on regularized regression for unsupervised learning of GRNs, with applications to data sets from different organisms. In addition, I will discuss findings from a recently developed approach based on network representation of transcriptomics data to learn GRNs in a supervised fashion. Emphasis will be placed on the accuracy of predictions and means for their improvements by considering other types of data (e.g. DAP-seq, ChIP-seq, motif binding). Finally, I will present preliminary results of applying these approaches to transcriptomics data from Chlamydomonas reinhardtii to discern factors contributing to regulating the response to high light.