Assessing insect biodiversity with automatic light traps in Brazil: Pearls and pitfalls of metabarcoding samples in...

Automated species identification based on data produced with metabarcoding offers an alternative for assessing biodiversity of bulk insect samples obtained with traps. We used a standard two-step PCR approach to amplify a 313 bp fragment of the barcoding region of the mitochondrial COI gene. The PCR products were sequenced on an Illumina MiSeq platform, and the OTUs production and taxonomic identifications were performed with a customized pipeline and database. The DNA used in the PCR procedures was extracted directly from the preservative ethanol of bulk insect samples obtained with automatic light traps in 12 sampling areas located in different biomes of Brazil, during wet and dry seasons. Agricultural field and forest edge habitats were collected for all sampling areas. A total of 119 insect OTUs and nine additional OTUs assigned to other arthropod taxa were obtained at a ≥97% sequence similarity level. The alpha and beta diversity analyses comparing biomes, habitats, and seasons were mostly inconclusive, except for a significant difference in beta diversity between biomes. In this study, we were able to metabarcode and HTS adult insects from their preservative medium. Notwithstanding, our results underrepresent the true magnitude of insect diversity expected from samples obtained with automatic light traps in Brazil. Although biological and technical factors might have impacted our results, measures to optimize and standardize eDNA HTS should be in place to improve taxonomic coverage of samples of unknown diversity and stored in suboptimal conditions, which is the case of most eDNA samples.