Background: Mitochondria contain small genomes that are physically separate from those of nuclei. Their comparison serves as a model system for understanding the processes of genome evolution. Although complete mitochondrial genome sequences have been reported for more than 600 animals, the taxonomic sampling is highly biased toward vertebrates and arthropods, leaving much of the diversity yet uncharacterized. Results: The mitochondrial genome of a cephalopod mollusk, the Chambered Nautilus, is 16,258 nts in length and 59.5 percent A+T, both values that are typical of animal mitochondrial genomes. It contains the 37 genes that are typical for animal mtDNAs, with 15 on one DNA strand and 22 on the other. The arrangement of these genes can be derived from that of the distantly related Katharina tunicata (Mollusca: Polyplacophora) by a switch in position of two large blocks of genes and transpositions of four tRNA genes. There is strong skew in the distribution of nucleotides between the two strands. There are an unusual number of non-coding regions and their function, if any, is not known; however, several of these demark abrupt shifts in nucleotide skew, suggesting that they may play roles in transcription and/or replication. One of the non-coding regions contains multiple repeats of a tRNA-like sequence. Some of the tRNA genes appear to overlap on the same strand, but this could be resolved if the polycistron were cleaved at the beginning of the downstream gene, followed by polyadenylation of the product of the upstream gene to form a fully paired structure. Conclusions: Nautilus sp. mtDNA contains an expected gene content that has experienced few rearrangements since the evolutionary split between cephalopods and polyplacophorans. It contains an unusual number of non-coding regions, especially considering that these otherwise often are generated by the same processes that produce gene rearrangements. This appears to be yet another case where polyadenylation of mitochondrial tRNAs restores what would otherwise bean incomplete structure.

Mitochondria and plastids (collectively called organelles) descended from prokaryotes that adopted an intracellular, endosymbiotic lifestyle within early eukaryotes. Comparisons of their remnant genomes address a wide variety of biological questions, especially when including the genomes of their prokaryotic relatives and the many genes transferred to the eukaryotic nucleus during the transitions from endosymbiont to organelle. The pace of producing complete organellar genome sequences now makes it unfeasible to do broad comparisons using the primary literature and, even if it were feasible, it is now becoming uncommon for journals to accept detailed descriptions of genome-level features. Unfortunately no database is currently useful for this task, since they have little standardization and are riddled with error. Here I outline what is currently wrong and what must be done to make this data useful to the scientific community.

Although the phylogenetic relationships of many organisms have been convincingly resolved by the comparisons of nucleotide or amino acid sequences, others have remained equivocal despite great effort. Now that large-scale genome sequencing projects are sampling many lineages, it is becoming feasible to compare large data sets of genome-level features and to develop this as a tool for phylogenetic reconstruction that has advantages over conventional sequence comparisons. Although it is unlikely that these will address a large number of evolutionary branch points across the broad tree of life due to the infeasibility of such sampling, they have great potential for convincingly resolving many critical, contested relationships for which no other data seems promising. However, it is important that we recognize potential pitfalls, establish reasonable standards for acceptance, and employ rigorous methodology to guard against a return to earlier days of scenario-driven evolutionary reconstructions.