Examining the utility of DNA barcodes for the identification of tallgrass prairie flora

Abstract

Premise: The tallgrass prairies of North America are one of the most threatened ecosystems in the world, making efficient species identification essential for understanding and managing diversity. Here, we assess DNA barcoding with high-throughput sequencing as a method for rapid plant species identification. Methods: Using herbarium collections representing the tallgrass prairie flora of Oak Lake Field Station, South Dakota, USA, we amplified and examined four common nuclear and plastid barcode regions (ITS, matK, psbA-trnH, and rbcL), individually and in combination, to test their success in identifying samples to family, genus, and species levels using BLAST searches of three databases of varying size. Results: Concatenated barcodes increased performance, although none were significantly different than single-region barcodes. The plastid region psbA-trnH performed significantly more poorly than the others, while barcodes containing ITS performed best. Database size significantly affected identification success at all three taxonomic levels. Confident species-level identification ranged from 8–44% for the global database, 13–56% for the regional database, and 21–80% for the sampled species database, depending on the barcode used. Discussion: Barcoding was generally successful in identifying tallgrass prairie genera and families, but was of limited use in species-level identifications. Database size was an important factor in successful plant identification. We discuss future directions and considerations for improving the performance of DNA barcoding in tallgrass prairies.