rbcL and matK earn two thumbs up as the core DNA barcode for ferns

DNA barcoding will revolutionize our understanding of fern ecology, most especially because the accurate identification of the independent but cryptic gametophyte phase of the fern's life history--an endeavor previously impossible--will finally be feasible. In this study, we assess the discriminatory power of the core plant DNA barcode (rbcL and matK), as well as alternatively proposed fern barcodes (trnH-psbA and trnL-F), across all major fern lineages. We also present plastid barcode data for two genera in the hyperdiverse polypod clade--Deparia (Woodsiaceae) and the Cheilanthes marginata group (currently being segregated as a new genus of Pteridaceae)--to further evaluate the resolving power of these loci.

Our results clearly demonstrate the value of matK data, previously unavailable in ferns because of difficulties in amplification due to a major rearrangement of the plastid genome. With its high sequence variation, matK complements rbcL to provide a two-locus barcode with strong resolving power. With sequence variation comparable to matK, trnL-F appears to be a suitable alternative barcode region in ferns, and perhaps should be added to the core barcode region if universal primer development for matK fails. In contrast, trnH-psbA shows dramatically reduced sequence variation for the majority of ferns. This is likely due to the translocation of this segment of the plastid genome into the inverted repeat regions, which are known to have a highly constrained substitution rate.

Our study provides the first endorsement of the two-locus barcode (rbcL+matK) in ferns, and favors trnL-F over trnH-psbA as a potential back-up locus. Future work should focus on gathering more fern matK sequence data to facilitate universal primer development.