Is LGT or Divergence a Better Explanation of Mitochondrial Genome Evolution?
To test the impact of LGT on the present mitochondrial data we examined the spectrum of phylogenetic signals for a given set of data and compared it to the spectrum obtained for a well-established phylogeny that contains sequences of different degrees of divergence. In doing this, phylogenetic signal could be expressed in many ways. A particularly convenient way is in terms of tree-splits (which are equivalent to edges, branches, or bipartitions). Internal splits separate OTUs (operational taxonomical units, in this case sequences) into two groups. External splits separate a single OTU from all other OTUs. Real data usually contain external splits and conflicting internal splits. For small numbers of OTUs, support for all splits under a specified model of sequence evolution can be calculated directly using a Hadamard transformation (Penny et al. 1996; Lockhart et al. 1999). For larger data sets, one heuristic approach is to use NNet, which provides a list of major splits, including con
