How and where to look for tRNAs in Metazoan mitochondrial genomes, and what you might find when you get there

The ability to locate and annotate mitochondrial genes is an important practical issue, given the rapidly increasing number of mitogenomes appearing in the public databases. Unfortunately, tRNA genes in Metazoan mitochondria have proved to be problematic because they often vary in number (genes missing or duplicated) and also in the secondary structure of the transcribed tRNAs (T or D arms missing). I have performed a series of comparative analyses of the tRNA genes of a broad range of Metazoan mitogenomes in order to address this issue. I conclude that no single computer program is necessarily capable of finding all of the tRNA genes in any given mitogenome, and that use of both the ARWEN and DOGMA programs is sometimes necessary because they produce complementary false negatives. There are apparently a very large number of erroneous annotations in the databased mitogenome sequences, including missed genes, wrongly annotated locations, false complements, and inconsistent criteria for assigning the 5' and 3' boundaries; and I have listed many of these. The extent of overlap between genes is often greatly exaggerated due to inconsistent annotations, although notable overlaps involving tRNAs are apparently real. Finally, three novel hypotheses were examined and found to have support from the comparative analyses: (1) some organisms have mitogenomic locations that simultaneously code for multiple tRNAs; (2) some organisms have mitogenomic locations that simultaneously code for tRNAs and proteins (but not rRNAs); and (3) one group of nematodes has several genes that code for tRNAs lacking both the D and T arms.