Quantification of marine benthic communities with metabarcoding

Abstract

DNA metabarcoding methods have been implemented in studies aimed at detecting and quantifying marine benthic biodiversity. In such surveys, universal barcodes are amplified and sequenced from environmental DNA. To quantify biodiversity with DNA metabarcoding, a relation between the number of DNA sequences of a species and its biomass and/or the abundance is required. However, this relationship is complicated by many factors, and it is often unknown. In this study, we validate estimates of biomass and abundance from molecular approaches with those from the traditional morphological approach. Abundance and biomass were quantified from 126 samples of benthic intertidal mudflat using traditional morphological approaches and compared with frequency of occurrence and relative read abundance estimates from a molecular approach. A relationship between biomass and relative read abundance was found for two widely dispersed annelid taxa (Pygospio and Scoloplos). None of the other taxons, however, showed such a relationship. We discuss how quantification of abundance and biomass using molecular approaches are hampered by the ecology of DNA i.e. all the processes that determine the amount of DNA in the environment, including the ecology of the benthic species as well as the compositional nature of sequencing data.

Keywords: abundance; biomass; eDNA; metabarcoding; next-generation sequencing; quantification.

FIGURE 1

Map of Texel, showing the sampling locations at the intertidal mudflats, NE of Texel. Also, a graphical display of the sampling scheme is shown. All points were sampled in 2016 at: 6 June and 14 November and in 2017: 13 March, 9 May, 23 May, 6 June and 26 June

FIGURE 2

Histogram showing frequency distribution of number of macrofauna taxa (genus) identified per sample for morphological (black) and molecular method (white). The grey area indicates overlap

FIGURE 3

Occurrence per taxon, calculated as the sum of detections divided by the total number of samples (n = 126) for both the morphological method (x‐axis) and the molecular method (y‐axis). Taxa detected belonged to three phyla: Annelida (red), Arthropoda (green) or Mollusca (blue)

FIGURE 4

Logistic regression and abundance distribution histograms for six annelid taxa. The histograms on the top show the abundance distribution if the species was also found in the molecular samples, the histograms on the bottom is the species was not present in the molecular data set, the frequency is shown on the y‐axis. The abundance distribution as derived from the morphological method are shown in histograms at the sqrt‐transformed x‐axis

FIGURE 5

Receiver operating characteristics (ROC) curve as derived from the logistic regression for six annelid taxa. The ROC curve shows the sensitivity on the y‐axis, calculated as the true positive predictions by the logistic model and the specificity, calculated as the false positive predictions, on the x‐axis

FIGURE 6

Relative read abundance approach relationships for six annelid taxa. The square‐root of biomass estimates as derived from the morphological method are shown on the x‐axis and the log₁₀ of the relative read abundance of the same sample in the molecular data set is shown on the y‐axis