Influence of pairwise genetic distance computation and reference sample size on the reliability of species identification using Cyt b and COI gene fragments in a group of native passerines.

Species identification is fundamental to wildlife forensic practice. The desirability of molecular genetic methods is increasing rapidly. The sequence of a marker, rather than its particular diagnostic nucleotides, provides greater safety through comparisons between intra- and inter-specific pairwise genetic distances. However, it has not been well described how reliability of species assignment is influenced by distance computing methods and reference sample sizes. In this study, the influences were tested using 12 species from 4 genera of passerine birds and the sequences of partial Cytochrome b (Cyt b) and Cytochrome Oxidase subunit I (COI) genes. Results showed that different substitution types have different outcomes of pairwise genetic distance estimation and this influences the risk of false inclusion and exclusion. Transition (Ts) is the most effective substitution type to reveal optimal species resolution for both Cyt b and COI gene fragments no matter whether K2P and p-distance are used. Sample size required to accurately estimate pairwise distance is essentially determined by the genetic diversity of a species in reference to a given strictness of predefined acceptable accuracy. These findings suggest that for future forensic work on birds by use of Cyt b and COI gene fragments, transition should be used exclusively for marker validation and identification practice when targeting closely related species. Meanwhile, the reference database should sufficiently represent overall genetic diversity of the species. The minimum sample size should be estimated based on existing knowledge of genetic diversity. Special caution should be used for species assignment when only several reference data are available for animals that are considered likely to have high genetic diversity.