Co-expression network analysis for identification of secondary metabolite pathways in Marchantia polymorpha

Abstract

The plant kingdom is home to a diverse collection of secondary metabolites with tremendous economic and therapeutic value. Secondary metabolite pathways are difficult to fully uncover for various reasons. Co-expression network analysis offers bioinformatic techniques and insights into fully elucidating these networks. The basal land plant Marchantia polymorpha’s genome containing limited genetic redundancy offers an exciting genetic window into the evolution of terrestrial land plants. In this project, co-expression networks of M. polymorpha were analysed to identify biological pathways, of which 85 pathways were found. While it is possible to identify biological pathways with this analysis, the predictive quality of the network still has room for improvement. Integration of enzyme commission numbers to the network, as a novel centrality-like edge weight, to identify true relationships was shown to improve network quality scores by up to 1.3 times. However, the predictive ability of the final network is still far from optimal, which highlights the need to address other possible limitations in the network, one possibility is the composition-bias of RNA-sequencing experiments.