Combining morpho-taxonomy and metabarcoding enhances the detection of non-indigenous marine pests in biofouling communities

Ulla Von Ammon,Anastasija Zaiko,Shane Lavery,Olivier Laroche,Graeme J Inglis,Xavier Pochon,Susanna A Wood,Leigh Tait

doi:10.1038/s41598-018-34541-1

Ulla Von Ammon, Anastasija Zaiko + Show 6 more

Open Access

https://doi.org/10.1038/s41598-018-34541-1

Copy DOI

Abstract

Marine infrastructure can favor the spread of non-indigenous marine biofouling species by providing a suitable habitat for them to proliferate. Cryptic organisms or those in early life stages can be difficult to distinguish by conventional morphological taxonomy. Molecular tools, such as metabarcoding, may improve their detection. In this study, the ability of morpho-taxonomy and metabarcoding (18S rRNA and COI) using three reference databases (PR2, BOLD and NCBI) to characterize biodiversity and detect non-indigenous species (NIS) in biofouling was compared on 60 passive samplers deployed over summer and winter in a New Zealand marina. Highest resolution of metazoan taxa was identified using 18S rRNA assigned to PR2. There were higher assignment rates to NCBI reference sequences, but poorer taxonomic identification. Using all methods, 48 potential NIS were identified. Metabarcoding detected the largest proportion of those NIS: 77% via 18S rRNA/PR2 and NCBI and 35% via COI/BOLD and NCBI. Morpho-taxonomy detected an additional 14% of all identified NIS comprising mainly of bryozoan taxa. The data highlight several on-going challenges, including: differential marker resolution, primer biases, incomplete sequence reference databases, and variations in bioinformatic pipelines. Combining morpho-taxonomy and molecular analysis methods will likely enhance the detection of NIS from complex biofouling.

Highlights

Biological invasions of non-indigenous species (NIS) can cause severe economic and environmental impact contributing to biodiversity loss[1,2]
The total paired-end, quality filtered and non-chimeric and de novo non-chimeric sequences obtained from combined summer and winter plates for 18S rRNA and COI Polymerase Chain Reaction (PCR) amplicons were 2,831,859 reads (95,751 unique Operational Taxonomic Units (OTUs)) and 3,084,751 reads (211,164 unique OTUs), respectively (Table 1)
Negative controls contained 41 reads assigned to 19 OTUs for 18S rRNA and COI, indicating insignificant background contamination, from which no new potential NIS were identified

Summary

Introduction

Biological invasions of non-indigenous species (NIS) can cause severe economic and environmental impact contributing to biodiversity loss[1,2]. GenBank[30] contains reference sequences from many different genetic markers and includes all domains of life but is prone to erroneously identified sequences[31]. Another limitation of metabarcoding is a lack of ‘universal’ markers across phyla[5], not all taxa can be detected due to primer selectivity and resulting amplification biases[32]. A 97% similarity threshold is commonly applied for general biodiversity assessments, a higher threshold may increase rare taxa detection, which may be well suited when targeting NIS36

Objectives

Methods

Results

Discussion

Conclusion