Identification of Birds through DNA Barcodes

Abstract

Short DNA sequences from a standardized region of the genome provide a DNA barcode for identifying species. Compiling a public library of DNA barcodes linked to named specimens could provide a new master key for identifying species, one whose power will rise with increased taxon coverage and with faster, cheaper sequencing. Recent work suggests that sequence diversity in a 648-bp region of the mitochondrial gene, cytochrome c oxidase I (COI), might serve as a DNA barcode for the identification of animal species. This study tested the effectiveness of a COI barcode in discriminating bird species, one of the largest and best-studied vertebrate groups. We determined COI barcodes for 260 species of North American birds and found that distinguishing species was generally straightforward. All species had a different COI barcode(s), and the differences between closely related species were, on average, 18 times higher than the differences within species. Our results identified four probable new species of North American birds, suggesting that a global survey will lead to the recognition of many additional bird species. The finding of large COI sequence differences between, as compared to small differences within, species confirms the effectiveness of COI barcodes for the identification of bird species. This result plus those from other groups of animals imply that a standard screening threshold of sequence difference (10x average intraspecific difference) could speed the discovery of new animal species. The growing evidence for the effectiveness of DNA barcodes as a basis for species identification supports an international exercise that has recently begun to assemble a comprehensive library of COI sequences linked to named specimens.

Figure 1. Comparison of Nucleotide Sequence Differences in COI among 260 Species of North American Birds

Pairwise comparisons between 437 COI sequences are separated into three categories: differences between individuals in the same species, differences between individuals in the same genus (not including intraspecific differences), and differences between individuals in the same family (not including intraspecific or intrageneric differences).

Figure 2. NJ Tree of COI Sequences from 30 Species in Family Scolapacidae (Sandpipers and Kin)

The divergent pair of clustered sequences of Tringa solitaria is highlighted. An asterisk indicates a COI sequence from GenBank.

Figure 3. NJ Tree of COI Sequences from 260 Species of North American Birds

Intraspecific divergences were sampled in 130 species; these are marked in blue. Four species showed deep intraspecific divergence: (a) Sturnella magna, (b) Cistothorus palustris, (c) Vireo gilvus, and (d) Tringa solitaria. Higher-order classifications in families (gray) and orders (gold) are highlighted, and are labeled on the left and right of the figure, respectively. Gold numerals indicate the two species that appear as paraphyletic lineages at the family level: (1) Oenanthe oenanthe and (2) Hirundo rustica.

Figure 4. Genetic Difference versus Geographic Distance

For each same-species pair of specimens, the geographic distance between where specimens were collected is plotted against their COI divergence (K2P).

Figure 5. Intraspecific Compared to Interspecific COI Distances (K2P) for Individual Species

For each species in which this comparison was possible (n = 73), maximum intraspecific variation is compared to minimum interspecific congeneric difference. For illustration purposes shown here, 2.0% is chosen as a cutoff between usual values for intra- and interspecific variation. This divides the graph into four quadrants that represent different categories of species: (I) Intraspecific distance, <2%; interspecific distance, >2%: concordant with current taxonomy; (II) Intraspecific distance, >2%; interspecific distance, >2%: probable composite species (i.e., candidate for taxonomic split); (III) Intraspecific distance, <2%; interspecific distance, <2%: recent divergence, hybridization, or synonymy; (IV) Intraspecific distance, >2%; interspecific distance, <2%: probable misidentification of specimen.