Reduced representation sequencing accurately quantifies relative abundance and reveals population-level variation in Pseudo-nitzschia spp

Abstract

Certain species within the genus Pseudo-nitzschia are able to produce the neurotoxin domoic acid (DA), which can cause illness in humans, mass-mortality of marine animals, and closure of commercial and recreational shellfisheries during toxic events. Understanding and forecasting blooms of these harmful species is a primary management goal. However, accurately predicting the onset and severity of bloom events remains difficult, in part because the underlying drivers of bloom formation have not been fully resolved. Furthermore, Pseudo-nitzschia species often co-occur, and recent work suggests that the genetic composition of a Pseudo-nitzschia bloom may be a better predictor of toxicity than prevailing environmental conditions. We developed a novel next-generation sequencing assay using restriction site-associated DNA (2b-RAD) genotyping and applied it to mock Pseudo-nitzschia communities generated by mixing cultures of different species in known abundances. On average, 94% of the variance in observed species abundance was explained by the expected abundance. In addition, the false positive rate was low (0.45% on average) and unrelated to read depth, and false negatives were never observed. Application of this method to environmental DNA samples collected during natural Pseudo-nitzschia spp. bloom events in Southern California revealed that increases in DA were associated with increases in the relative abundance of P. australis. Although the absolute correlation across time-points was weak, an independent species fingerprinting assay (Automated Ribosomal Intergenic Spacer Analysis) supported this and identified other potentially toxic species. Finally, we assessed population-level genomic variation by mining SNPs from the environmental 2bRAD dataset. Consistent shifts in allele frequencies in P. pungens and P. subpacifica were detected between high and low DA years, suggesting that different intraspecific variants may be associated with prevailing environmental conditions or the presence of DA. Taken together, this method presents a potentially cost-effective and high-throughput approach for studies aiming to evaluate both population and species dynamics in mixed samples.

Keywords: 2b-RAD; Community composition; Harmful algal bloom; Next-generation sequencing; Population genomics.

Figure 1.

False positive rate (FPR) relative to high quality mapped reads as a function of read depth (a) and whether the sample consisted of a single species or multi-species pool (b). Individual samples are colored according to their mix type ratio (Table 1).

Fig. 2.

Accuracy of a reduced representation sequencing (2bRAD) method for quantifying species relative abundance. (a) Observed abundance, calculated as the proportion of reads mapping to the reference library, as a function of expected abundance based on the target proportion in the mock community mix varies by species. The dashed gray line is the 1:1 line. (b) Accuracy, or the deviation between observed and expected values, as a function of species, expected abundance of the focal sample (see legend at right) and mix ratio (see legend at right). (c) Accuracy as a function of mix ratio, expected abundance of the focal sample (see legend at right) expected abundance of the focal sample (see legend at right) and species. (d) Accuracy as a function of source culture (denoted with alpha-numeric labels), expected abundance of the focal sample (see legend at right) and mix ratio (see legend at right).

Figure 3.

Dynamics of Pseudo-nitzschia spp. blooms sampled from Newport Beach Pier, CA. (a-d) The concentration of particulate domoic acid (ng/mL) in seawater. (e-h) The concentration of small size class (< 3 µm) cells per liter. (i-l) The concentration of large size class (> 3 µm) cells per liter. (m-p) The percent of high quality reads exhibiting high quality mapping to the Pseudo-nitzschia spp. reference library. (q-t) The relative abundance of focal Pseudo-nitzschia spp. calculated as the proportion of high quality reads exhibiting high quality mapping to each individual species relative to the total number of reads mapping to the Pseudo-nitzschia spp. reference library.

Fig. 4.

ARISA-based diversity profile of Pseudo-nitzschia spp. blooms sampled from Newport Beach Pier, CA in (a) 2015, (b) 2017, (c) 2018, and (d) 2019. Relative abundance data (based on% of total relative fluorescence units in the ARISA) shown for all 17 amplicons detected. Species are defined by name (if known) and amplicon size in base pairs. * samples with low total peak height.

Figure 5.

Minor allele frequency at select loci (CMH<0.05) by sampling year, stratified by the relative amount of domoic acid detected over the course of the sampling window. Ppun = P. pungens, Psub = P. subpacifica.