Community Structure in Transcriptional Regulatory Networks of Yeast Species

The genomic expression of living organisms is controlled by complex transcriptional regulation. Transcriptional regulatory networks, composed by associations between transcription factors and target genes, are responsible for representing and controlling this gene expression and regulate the response of an organism to environmental changes. In this paper, we extend the study of these systems by applying different community detection algorithms on closely-related yeast transcriptional regulation networks to characterize their topological structure and understand if these methods are able to capture meaningful functional clusters of genes. We start by evaluating the accuracy and efficiency of a large group of algorithms by applying them to benchmark networks with ground-truth communities. We then apply the methods that had the best performance to the yeast networks to analyze the quality of the resulting structures, and then, assess the quality of the retrieved modules from a biological point of view using available annotated species' biological functions. Finally, we apply a multilayer community detection algorithm on multilayer networks, where each layer is an individual yeast network, and use available mappings between nodes of different species to successfully discover communities with genes that belong to different species but have similar biological functions. We conclude that the use of community detection algorithms to functionally characterize the modules of these networks might not be enough, suggesting the need of additional genetic information and possibly the use of alternative strategies to study these complex regulatory networks.