How reliable is mitochondrial DNA for recovering the phylogeny of Chironomidae (Culicomorpha: Diptera)?

We examine mitochondrial DNA as a source of data to estimate phylogenetic relationships within the Chironomidae (Diptera). Previous studies have shown that mitogenomes often produce ambiguous phylogenetic topologies that are inconsistent with both morphological and multi-locus molecular data. In this study, we sequenced 18 new mitogenomes representing 5 subfamilies, including the first available sequence for Protanypodinae. These were combined with 65 previously annotated chironomids mitogenomes, 8 additional individuals assembled from SRA data, and 8 outgroup taxa. Phylogenetic reconstructions were performed using complete protein-coding genes (PCGs), the first and second codon positions of PCGs (PCG 12), with and without ribosomal genes (12S rDNA and 16S rDNA), and amino acid sequences (AA). Both Bayesian Inference and Maximum Likelihood approaches were implemented. Three alternative outgroup compositions were tested: (i) chironomids only, with restricted rooting on Podonominae; (ii) Ceratopogonidae and Culicidae; and (iii) a diverse selection of Culicomorpha. We found that the AA, PCG12, and PCG 12 + rDNA datasets, when coupled with the third outgroup combination, provide the strongest phylogenetic signal, with the highest effective sample size and log-likelihood scores. In most cases, the resulting tree topologieswere congruent between mitochondrial and multi-locus data. However, some consistent differences in topologies were observed, leading to differences in divergence time estimates. Our phylogenetic study indicates paraphyly of Orthocladiinae due to the positions of Brillia Kieffer and Abiskomyia Edwards, suggesting that a comprehensive integrative revision of this subfamily is required. We conclude that the reliability of the mitochondrial phylogenetic signal improves with the increased taxon sampling.