PS 42-91
DNA barcoding a complex prairie flora using the rbcL-matK gene regions

Wednesday, August 7, 2013
Exhibit Hall B, Minneapolis Convention Center
John A. Barone, Biology, Columbus State University, Columbus, GA
Robert B. Futrell, Biology, Columbus State University, Columbus, GA
Kevin S. Burgess, Biology, Columbus State University, Columbus, GA
Background/Question/Methods

DNA barcoding is a molecular technique that uses diverse gene regions to identify and distinguish even closely related species.  In animal DNA barcoding, the CoI gene region of the mitochondrial genome has become the standard barcode.  For plants, a number of gene regions have been proposed as possible DNA barcodes, including the chloroplast gene regions, rbcL and matK.  Used in conjunction, these two regions have yielded relatively high species discrimination at a relatively low cost in previous studies.  For example, the CBOL Plant Working Group (2009) found that rbcL- matK barcode yielded 72% resolution for plant samples across a broad taxonomic range, including gymnosperms and cryptograms.  Burgess et al. (2011) had 93.1% resolution for plant samples from a more moderate taxonomic range.  In this project, we have been examining the effectiveness of rbcL-matK in distinguishing and identifying species from a complex prairie flora that includes a high proportion of polytypic genera and a narrow taxonomic range.  We expected that the effectiveness of rbcL and matK would drop for this type of flora.  Leaf samples were taken from herbarium specimens that had been collected from 31 prairie sites from the Black Belt region of Alabama and Mississippi from 2005 to 2012.  DNA isolation and sequencing was done by BOLD Systems.  DNA sequences were edited and compared with those in BOLD Systems and Genbank. 

Results/Conclusions

To date, we have barcoded 190 samples from 80 herbaceous species.  Both regions, rbcL and matK, were recovered in 92.6% of the samples.  Monotypic genera had a higher proportion of species resolution (100%) than polytypic genera (66.7%).  There was a significant, negative correlation between the number of species per genera and percent species resolution (Spearman’s rho = -0.817, P<0.001). These results suggest that for floras with a high number of closely related species, rbcL and matK have reduced ability to distinguish species.  Additional gene regions, or intergenic spacers, such as trnH-psbA, may be needed to supplement rbcL and matK in these situations.