Genetic Code Prediction for Metazoan Mitochondria with GenDecoder

There is a standard genetic code that is used by most organisms, but exceptions exist in which particular codons are translated with a different meaning, i.e., as a different amino acid. The characterization of the genetic code of an organism is hence a key step for properly analyzing and translating its protein-coding genes. Such characterization is particularly important in the case of metazoan mitochondrial genomes for two reasons: first, many variant codes occur in them and second, mitochondrial data is frequently used for evolutionary studies. Variant codes are usually found by comparative sequence analyses. Given a protein alignment, if a particular codon for a given species occurs at positions in which a particular amino acid is frequently found in other species, then the most likely hypothesis is that the codon is translated as that particular amino acid in that species. Previously, we have shown that this method can be very reliable provided that enough taxa and positions are included in the comparisons and have implemented it in the web-ser GenDecoder (http://darwin.uvigo.es/software/gendecoder.html). In this chapter we describe the rationale of the method used by GenDecoder and its usage through worked examples, highlighting the potential problems that can arise during the analysis.