Between a Rock and a Hard Polytomy: Phylogenomics of the Rock-Dwelling Mbuna Cichlids of Lake Malaŵi

Whole genome sequences are beginning to revolutionize our understanding of phylogenetic relationships. Yet, even whole genome sequences can fail to resolve the evolutionary history of the most rapidly radiating lineages, where incomplete lineage sorting, standing genetic variation, introgression, and other factors obscure the phylogenetic history of the group. To overcome such challenges, one emerging strategy is to integrate results across different methods. Most such approaches have been implemented on reduced representation genomic data sets, but whole genomes should provide the maximum possible evidence approach. Here, we test the ability of single nucleotide polymorphisms extracted from whole genome resequencing data, implemented in an integrative genomic approach, to resolve key nodes in the phylogeny of the mbuna, rock-dwelling cichlid fishes of Lake Malaŵi, which epitomize the phylogenetic intractability that often accompanies explosive lineage diversification. This monophyletic radiation has diversified at an unparalleled rate into several hundred species in less than 2 million years. Using an array of phylogenomic methods, we consistently recovered four major clades of mbuna, but a large basal polytomy among them. Although introgression between clades apparently contributed to the challenge of phylogenetic reconstruction, reduction of the data set to nonintrogressed sites still did not help to resolve the basal polytomy. On the other hand, relationships among six congeneric species pairs were resolved without ambiguity, even in one case where existing data led us to predict that resolution would be difficult. We conclude that the bursts of diversification at the earliest stages of the mbuna radiation may be phylogenetically unresolvable, but other regions of the tree are phylogenetically clearly supported. Integration of multiple phylogenomic approaches will continue to increase confidence in relationships inferred from these and other whole-genome data sets. [Incomplete lineage sorting; introgression; linkage disequilibrium; multispecies coalescence; rapid radiation; soft polytomy.]