Towards quantitative metabarcoding of eukaryotic plankton: an approach to improve 18S rRNA gene copy number bias

Abstract

Plankton metabarcoding is increasingly implemented in marine ecosystem assessments and is more cost-efficient and less time-consuming than monitoring based on microscopy (morphological). 18S rRNA gene is the most widely used marker for groups’ and species’ detection and classification within marine eukaryotic microorganisms. These datasets have commonly relied on the acquisition of organismal abundances directly from the number of DNA sequences (i.e. reads). Besides the inherent technical biases in metabarcoding, the largely varying 18S rRNA gene copy numbers (GCN) among marine protists (ranging from tens to thousands) is one of the most important biological biases for species quantification. In this work, we present a gene copy number correction factor (CF) for four marine planktonic groups: Bacillariophyta, Dinoflagellata, Ciliophora miscellaneous and flagellated cells. On the basis of the theoretical assumption that ‘1 read’ is equivalent to ‘1 GCN’, we used the GCN median values per plankton group to calculate the corrected cell number and biomass relative abundances. The species-specific absolute GCN per cell were obtained from various studies published in the literature. We contributed to the development of a species-specific 18S rRNA GCN database proposed by previous authors. To assess the efficiency of the correction factor we compared the metabarcoding, morphological and corrected relative abundances (in cell number and biomass) of 15 surface water samples collected in the Belgian Coastal Zone. Results showed that the application of the correction factor over metabarcoding results enables us to significantly improve the estimates of cell abundances for Dinoflagellata, Ciliophora and flagellated cells, but not for Bacillariophyta. This is likely to due to large biovolume plasticity in diatoms not corresponding to genome size and gene copy numbers. C-biomass relative abundance estimations directly from amplicon reads were only improved for Dinoflagellata and Ciliophora. The method is still facing biases related to the low number of species GCN assessed. Nevertheless, the increase of species in the GCN database may lead to the refinement of the proposed correction factor.

Metabarcoding vs Microscopy: Comparison of Methods To Monitor Phytoplankton Communities

“…43 However, there is no database with the number of gene copy per cell, and little information is available, i.e., few species in few environments, resulting in very coarse conversion factors per taxonomic groups, and no significant improvement in the comparison between metabarcoding and microscopy. 44 We found relative carbon biomass to better match the gene relative abundance than biovolume. This can be explained by the carbon conversion used in this study, 9 which takes into account the specificity of diatoms.…”

“…To overcome such a range of size variation, metabarcoding data have been suggested to be corrected by the number of gene copy per cell . However, there is no database with the number of gene copy per cell, and little information is available, i.e., few species in few environments, resulting in very coarse conversion factors per taxonomic groups, and no significant improvement in the comparison between metabarcoding and microscopy …”

Ecological divergence of a mesocosm in an eastern boundary upwelling system assessed with multi-marker environmental DNA metabarcoding

“…A. sanguinea reads were a minor proportion of the reads in the COI dataset and absent from the 16S rRNA plastid sequences; the disparity in the detection of A. sanguinea by our different PCR primers can be explained by inefficient dinoflagellate amplification by COI primers (Lin et al, 2009) and the loss of most chloroplast genes from dinoflagellates (Koumandou et al, 2004), causing them not to be detected in our 16S rRNA plastidial sequences (Needham and Fuhrman, 2016). While the correlation between cell biomass and amplicon sequence abundance is complicated both by technical biases in metabarcoding and variation in gene copy numbers (Martin et al, 2022), the high biomass of certain taxa as inferred using read numbers is corroborated by other sampling techniques used to sample community composition. In this experiment, A. sanguinea was also identified as the dominant dinoflagellate by imaging flow cytometry and microscopy in mesocosm M1 and in seven of eight mesocosms overall (Bach et al, 2020).…”

Relative importance of bacterivorous mixotrophs in an estuary-coast environment

“…This, along with biases in the amplification reaction (Bradley et al, 2016; Gonzalez et al 2012), can result in relative gene abundances in amplicon pools that are significantly different from cell abundances (Karlusich et al, 2022). To correct for this bias, we followed the principle of a study by Martin et al (2022) and developed our own empirical gene-to-cell correction factors (C.F.) for seven major groups that were distinguishable by morphology under microscopy (dinoflagellates, ciliates, radiolaria, cryptophytes, diatoms, other pigmented and heterotrophic eukaryotes), which were formulated as: Where %MB_gene g and %EM_cell g are the relative abundances of group g derived from metabarcoding 18S rRNA reads, and from epifluorescence microscopy-based cell abundances, respectively.…”