After extracting DNA from the whole specimen using non-destructive techniques, the insects are prepared dry for their subsequent identification and study. Total DNA is used as template for PCR amplification of an animal genetic marker (a fragment of the mtDNA gene for the subunit 1 of the cytochrome c oxidase, cox1) and a plant genetic marker (the cpDNA intron between the genes psbA and trnH).
The cox1 sequences are used to delimit leaf beetle species in the sample using a phylogenetic method which groups samples within evolutionary lineages with a diversification pattern compatible with population processes, therefore as a single species. This method is known as GMYC, the generalized mixed Yule-coalescent model.
On the other hand, one or several psbA-trnH sequences obtained from each beetle specimen are identified by comparison with taxonomically identified sequences available in public nucleotide sequence databases. Taxonomic assignment exploits a combination of phenetic (by similarity with GenBank sequences using BLAST-like searches) and phylogenetic (by phylogenetic placement of diet sequences within a phylogenetic tree with related plant sequences) methods.
The taxonomic coverage of sequences available in GenBank is fragmentary, and taxonomic assignment is facilitated in this project using a Nicaraguan plant reference sequence database. Thus, we have sampled as much plant diversity as possible from the same places were we have captured insects, and all plant species which could be identified were sequence for the same psbA-trnH marker as used for diet characterization. The availability of this reference database facilitates the taxonomic identification of diets, sometimes down to the species level (whereas taxonomic assignment based on GenBank only is usually restricted to plant family, tribe, or genus in some cases).