Construction of a COI DNA barcode reference library for non-volant terrestrial mammals of Taiwan, with notes on the effects of nuclear mitochondrial pseudogenes (numts) on molecular identification
TL;DR
The utility of COI barcodes for identifying most non-volant terrestrial mammals in Taiwan, and the importance of integrating specimen traceability, functional sequence checks, comparisons with public data, distance-based analyses, and cautious interpretation of BIN results in mammalian DNA barcoding research are emphasized.
Abstract
Abstract DNA barcoding using mitochondrial COI is widely applied for animal species identification, but its accuracy in mammals can be affected by uneven sampling, sequence artefacts, and nuclear mitochondrial pseudogenes (numts). In this study, we established a voucher-linked COI DNA barcode reference library for non-volant terrestrial mammals of Taiwan and evaluated its identification performance and potential limitations using the BOLD workbench, Barcode Index Number (BIN) assignments, barcode gap analyses, and comparisons with public sequence data. A total of 157 newly generated COI sequences were obtained, representing seven orders, 16 families, 37 genera and 48 species. Among the 48 sampled taxa, 15 (31.25%) had no COI barcode record available in BOLD before this study; therefore, the deposition of the present library increasing BOLD coverage for the sampled taxa from 68.8% to 100%. In the original BOLD project dataset, 145 of the 157 sequences were assigned to 49 BINs. To evaluate whether the observed barcode patterns were robust for broader sampling, 75 public Taiwanese mammalian COI sequences were incorporated into an expanded Taiwan dataset of 232 sequences. Barcode-gap and Neighbor-Joining analyses showed that 45 of the 48 taxa retained positive barcode gaps after public sequences were added, supporting the utility of COI barcodes for identifying most non-volant terrestrial mammals in Taiwan. Three taxa, Mogera insularis, M. kanoana, and Petaurista grandis, showed negative barcode gaps and/or split clustering patterns and were therefore highlighted for further investigation. In addition, all three newly generated Rattus tanezumi sequences were flagged as putative numts based on anomalous placement and full-length sequence inspection. Our results further emphasize the importance of integrating specimen traceability, functional sequence checks, comparisons with public data, distance-based analyses, and cautious interpretation of BIN results in mammalian DNA barcoding research.