You are here:
Sensitivity and accuracy of high-throughput metabarcoding methods for early detection of invasive fish species
Citation:
Hatzenbuhler, C., J. Kelly, J. Martinson, S. Okum, AND E. Pilgrim. Sensitivity and accuracy of high-throughput metabarcoding methods for early detection of invasive fish species. Scientific Reports. Nature Publishing Group, London, Uk, 7:1-10 (46393), (2017).
Impact/Purpose:
For early detection biomonitoring of aquatic invasive species, the ability to find rare individuals is critical and requires accurate species level identification to ensure all species present are accounted for. However, traditional taxonomic identification of many aquatic organisms requires substantial expertise and yet not all individuals may be accurately identified to species level. Advanced DNA sequencing techniques have been rapidly gaining recognition as an alternative means to identify species in complex environmental samples. To improve our understanding of the usefulness of new DNA based methods for early detection biomonitoring, we conducted several experiments to test the sensitivity and accuracy of these methods. Our findings confirmed the levels of detection possible with the processing methods used and identified several factors that influenced overall accuracy of the DNA based measurement of sample biodiversity.
Description:
For early detection biomonitoring of aquatic invasive species, sensitivity to rare individuals and accurate, high-resolution taxonomic classification are critical to minimize detection errors. Given the great expense and effort associated with morphological identification of many aquatic organisms, high-throughput DNA metabarcoding has gained recognition as an alternative means to describe biodiversity in aquatic community samples. DNA based techniques are advancing, but, our understanding of the limits to detection, including sensitivity to rare organisms and accuracy of results for metabarcoding complex samples is inadequate. We carried out several experiments designed to investigate the sensitivity and accuracy of metabarcoding methods commonly used to characterize sample composition. Multi-species assemblages were constructed using larval fish tissue from several “common” species and spiked with varying proportions of tissue from an additional “rare” target species. Pyrosequencing of the genetic marker, COI (cytochrome c oxidase subunit I) and subsequent sequence data analysis provided convincing experimental evidence of low-level detection for some species in cases where their relative biomass percentages were as low as 0.02% of total sample mass, but the limits to detection vary interspecifcally. Our experiments demonstrated that detection accuracy was susceptible to several factors that skewed sequence-based biodiversity measurements from the known relative biomass abundances, and increased the rate of false absence detection errors. COI amplification bias and sequence data processing methods were the main sources of error. We urge caution in interpreting presence/absence and relative abundance in multi-species larval fish assemblages until metabarcoding methods can be optimized for accuracy and precision.
