Reconciliation of Conflicting Phenotypic and rRNA Gene Phylogenies of Fungi in Glomeromycota Based on Underlying Patterns and Processes

Abstract

A true species phylogeny requires congruence amongst all independent character data sets that inform on continuity of descent. A morphology-based phylogeny conflicts with rRNA phylogenies in basal nodes, but all are in greater accord at the family level and above. Incongruent patterns are attributed to population-level processes in gene evolution that create signals of homoplasy and distorted relationships widely recognized as problematic in other higher eukaryotic lineages. A species phylogeny can be misrepresented by two processes, either separately or together. Prerequisites for both are: (1) cladogenesis of a gene occurs prior to species cladogenesis, and (2) polymorphisms from gene cladogenesis are preserved in both diverging species lineages and then are sorted by selection or loss. Invoking the former process, polymorphisms arise in rRNA gene loci on different chromosomes and therefore are not homogenized by concerted evolution. Conflicts arise biologically, when extinction or sorting of these polymorphisms do not correspond with other characters indicative of speciation, or methodologically when intra-isolate clones are undersampled. With the latter process, polymorphisms arise at a single multicopy locus, after which sorting of lineages in subsequent cladogenic events can result in orthologous sequences in one lineage and paralogous sequences in another. A species phylogeny will be misleading when sequences from different fungal isolates of a species are undersampled. When either process is considered, the discord between rRNA gene and morphology-based trees can be explained. There is no definitive data to support either process, but credence is reflected in the complete lack of support for the proposed rRNA gene phylogeny from morphological, biochemical, and ecological characters correctly assessed as homologs. New gene trees can clarify species evolution at the molecular level, but they also can be in conflict. When this occurs, data from all classes of characters must be analyzed within the framework of a balanced multidisciplinary approach.