Reconsidering cytonuclear discordance in the genomic age

Drew Allen Larson , Michael Itgen, Robert Denton, Matthew W Hahn

Historically, phylogenetic datasets had relatively few loci but were overrepresented for cytoplasmic sequences (mitochondria and chloroplast) because of their ease of amplification and large numbers of informative sites. Under those circumstances, it made sense to contrast individual gene tree topologies obtained from cytoplasmic loci and nuclear loci, with the goal of detecting differences between them—so-called cytonuclear discordance. In the current age of phylogenomics and ubiquitous gene tree discordance among thousands of loci, it is important to distinguish between simply observing discordance between cytoplasmic trees and a species tree inferred from many nuclear loci and identifying the cause of discordance. Here, we examine what inferences one can make from trees representing different genomic compartments. While topological discordance can be caused by multiple factors, the end goal of many studies is to determine whether the compartments have different evolutionary histories: what we refer to as “cytonuclear dissonance.” Answering this question is more complex than simply asking whether there is discordance, requiring additional analyses to determine whether genetic exchange has affected only (or mostly) one compartment. Furthermore, even when these histories differ, expectations about why they differ are not always clear. We conclude by pointing to current research and future opportunities that may help to shed light on topological variation across the multiple genomes contained within a single eukaryotic cell.