Exploring phylogenetic informativeness and nuclear copies of mitochondrial DNA (numts) in three commonly used mitochondrial genes: mitochondrial phylogeny of peppermint,cleaner, and semi‐terrestrial shrimps (Caridea: Lysmata,Exhippolysmata, and Merguia)

Of paramount importance to studies that profit from molecular trees is the accuracy and robustness of the reconstructed phylogenies. Causes of systematic error that can mislead phylogenetic methods include nuclear copies of mitochondrial DNA (numts) and low phylogenetic informativeness (PI). Herein, numts and PI were explored in three mitochondrial genes commonly used for phylogenetic reconstruction: 16S, 12S, and cytochrome c oxidase I (COI). Shrimps from the genera Lysmata, Exhippolysmata, and Merguia were used as a model system. The existence of: (1) multiple bands on gels of COI and 12S polymerase chain reaction (PCR) products from various species; (2) double peaks, background noise, and ambiguity in sequence chromatograms of COI and 12S PCR products that produced a single clear band in other species; and (3) indels, stop codons, and considerable composition bias in COI‐like cloned sequences of one problematic species (Lysmata seticaudata), was interpreted as evidence of pervasive non‐functional nuclear copies of mitochondrial DNA (numts) of the targeted COI (and probably 12S) mtDNA fragment. The information content of the three mtDNA markers studied was investigated using PI profiling, spectral analysis, and neighbour‐nets. Marker‐specific PI profiles suggested that the COI marker has the highest information content and greatest power for resolving both shallow and deep nodes in trees depicting the phylogenetic relationship among the species studied. Nonetheless, spectral analysis of splits and neighbour‐nets suggested that the 16S and 12S markers were equally or even more powerful than the COI marker for resolving nodes at all phylogenetic levels. Altogether, these analyses suggest that all three mtDNA markers are equally useful for resolving phylogenetic relationships in the shrimps studied, and that PI profiling is not necessarily useful to estimate overall gene utility. A ‘total‐evidence’ phylogenetic analysis that included 34 species and used a concatenated data set of 1403 characters (from reliable 16S, 12S and COI sequences), demonstrated that the genus Lysmata is paraphyletic, and that the monophyletic clade comprising species of Lysmata and Exhippolysmata can be divided into four well‐supported subclades (Neotropical, Cleaner, Cosmopolitan, and Morphovariable). © 2013 The Linnean Society of London