Optimization and performance testing of a sequence processing pipeline applied to detection of nonindigenous species

Abstract

Genetic taxonomic assignment can be more sensitive than morphological taxonomic assignment, particularly for small, cryptic or rare species. Sequence processing is essential to taxonomic assignment, but can also produce errors because optimal parameters are not known a priori. Here, we explored how sequence processing parameters influence taxonomic assignment of 18S sequences from bulk zooplankton samples produced by 454 pyrosequencing. We optimized a sequence processing pipeline for two common research goals, estimation of species richness and early detection of aquatic invasive species (AIS), and then tested most optimal models’ performances through simulations. We tested 1,050 parameter sets on 18S sequences from 20 AIS to determine optimal parameters for each research goal. We tested optimized pipelines’ performances (detectability and sensitivity) by computationally inoculating sequences of 20 AIS into ten bulk zooplankton samples from ports across Canada. We found that optimal parameter selection generally depends on the research goal. However, regardless of research goal, we found that metazoan 18S sequences produced by 454 pyrosequencing should be trimmed to 375–400 bp and sequence quality filtering should be relaxed (1.5 ≤ maximum expected error ≤ 3.0, Phred score = 10). Clustering and denoising were only viable for estimating species richness, because these processing steps made some species undetectable at low sequence abundances which would not be useful for early detection of AIS. With parameter sets optimized for early detection of AIS, 90% of AIS were detected with fewer than 11 target sequences, regardless of whether clustering or denoising was used. Despite developments in next‐generation sequencing, sequence processing remains an important issue owing to difficulties in balancing false‐positive and false‐negative errors in metabarcoding data.

Uses and Misuses of Environmental DNA in Biodiversity Science and Conservation

“…A suite of programs facilitates the identification and removal of both PCR and sequencing errors (reviewed in Coissac et al 2012). Thus, sequence processing is often performed to reduce technical errors, but the bioinformatics steps can be an important source of error because appropriate parameters are often not known (Clare et al 2016, Scott et al 2018. Flynn et al (2015) evaluated the performance of commonly used clustering methods on a series of mock communities and on more complex natural communities of zooplankton.…”

“…The performance of bioinformatics pipelines can be tested through simulation. For example, Scott et al (2018) tested 1,050 combinations of parameters to determine the optimal parameter sets for particular research goals. They then tested the pipeline performance (detectability and sensitivity) by computationally inoculating sequences from 20 aquatic invasive species into 10 zooplankton community samples, revealing that optimal parameter selection often depends on the research goal.…”

Early Alert of Biological Risk in a Coastal Lagoon Through eDNA Metabarcoding

“…OTU tables, a list of OTUs obtained for each sample and the number of sequences assigned to them, were constructed, clustering reads with a 100% identity between them and maintaining all assigned sequences including singletons to retain maximum sensitivity for species detection. The removal of singletons is usually employed to eliminate false positives as proposed by Scott et al (2018), in the context of species survival, NIS early detection, or marine biosecurity surveillance; a false negative is most costly than a false positive (von Ammon et al, 2018). Sequences of organisms without relevance for the study (e.g., human, insects, terrestrial plants, etc.)…”

Status and prospects of marine NIS detection and monitoring through (e)DNA metabarcoding

“…reference databases selection for sequences comparison) (Zhan et al 2014, Flynn et al 2015. In our review, we found that different pipelines have been employed in the bioinformatics workflows (Hatzenbuhler et al 2017, Scott et al 2018, von Ammon et al 2018b (Table S5), which further complicates comparison among studies.…”

“…However, similarity thresholds can be strongly dependent on the marker and the length of the fragments targeted. While for the 20 18S rRNA gene (more conserved) an increase from 97% to 99% OTUs clustering may allow species to be split, for typical COI fragments, which display high sequence variability, this alteration may not produce any meaningful increase in the number of species detected (Scott et al 2018). The same applies with the length of the fragments under analysis, with shorter fragments being probably more sensitive to similarity thresholds than longer fragments.…”