Molecular species identification and phylogenetic assignments of Bursearceae from Indonesia using plant core DNA barcoding

Species identification in tropical forests with high diversity and limited floral information is often difficult. In the last decade, to discriminate species and facilitate biodiversity studies, DNA barcoding has been applied in tropical forests such as Sumatran forests, which face high deforestation rates. DNA barcoding is a technique that uses short DNA sequences to discriminate species. The technique can help to identify morphologically similar species and overcome the limitation of traditional species identification which depends on short lived reproductive parts. A universal DNA barcode for animals has been identified but a universally suitable barcode for plants is still yet to be agreed upon. For their universality and discriminatory power, matK and rbcL are the most accepted barcode markers for plants.
The objectives of this study were to evaluate the efficiency of rbcL, matK and their combination for species identification, to estimate a species-tree inference and to generate DNA barcodes of species of family Burseraceae from Sumatra, Indonesia. For this study, 197 specimens representing 17 species from 4 genera of the family were collected. At the species level matK identified the highest percentage of specimens (25%) followed by matK+rbcL (22%) and rbcL (5%) using a BLAST search. However, based on the interspecific divergence, matK could distinguish 96% of the species. With a mean of 0.0009, matK also had the highest interspecific divergence. Besides, matK has 0.000 mean intraspecific divergences, whereas rbcL and the combined dataset have 0.0004. In addition to that, a barcode gap was only found in matK and matK+rbcL. Furthermore, molecular species identification was inferred by employing species-tree inferences using Maximum likelihood and Bayesian inference methods. The matK and matK+rbcL barcodes resolved 43% and 50% of the species as monophyletic using Maximum likelihood and Bayesian inference methods, respectively.