Introduction

Understanding ecosystems require construction and modelling of complex networks that represent various species interactions, and abiotic factors. In such ecosystems, small invertebrates form important components as consumers, prey, decomposers, pollinators and ecosystem engineers, and are thus critical to include (Kellert, 1993), but large and charismatic animals like birds and mammals have to date garnered the bulk of trophic ecologists attention (Eisenhauer & Hines, 2021). Traditionally, dietary studies have been conducted through direct observation of feeding behavior, or morphological identification of prey species in regurgitate, stomach or fecal contents (Pompanon et al., 2012; Sousa et al., 2019; Symondson, 2002), but these approaches are highly impractical for a large portion of invertebrates whose small bodies complicate stomach content extraction and produce comparatively small fecalia.
Microscopy has helped identify small prey from small consumers, but is laborious and biased in favor of big, well-preserved prey (Berg, 1979), and demands morphological and taxonomic expertise (Pompanon et al., 2012). Conversely, experimental monitoring of communities over time can detect trophic interactions, and even allow quantifying ingestion rates, but struggles in reproducing the natural variability. The species assortment would typically be limited to prey expected in advance, or that co-occurred with the consumer if a natural sample was used as the starting point. Prey metabarcoding has become popular because it allows identification of diverse prey from complex and partly digested material, and does not require considerable a priori knowledge of prey, or taxonomical or morphological knowledge from the researcher (Casper et al., 2007). In broad strokes, metabarcoding includes extraction of DNA from dietary material – most often regurgitate, feces or stomach content, PCR amplification of target DNA (the marker gene or barcode), sequencing of PCR amplicons, and culminates with taxonomic identification by comparing obtained sequences to those in a reference database (Santoferrara, 2019).
Deciding on a dietary material is an important step that depends on the logistics and ethics of sampling, or the nature of the species being studied, such as its size or tendency for violence (Pompanon et al., 2012). Medium-sized crustaceans like northern shrimp (Pandalus borealis ) may be suitable for excision of stomach content (Urban et al., 2022), but small invertebrates (< 1 mm) are challenging to dissect, and may require whole body extraction (e.g. Novotny et al., 2021; Zamora-Terol et al., 2020). This comes at a cost, however, because the majority of DNA in the sample will naturally stem from the consumer itself (Piñol et al., 2014, 2015). An overabundance of consumer DNA may also be a challenge when other materials (e.g., feces or gut content) are sampled (Kohn & Wayne, 1997), but its concentration becomes severely exalted in extracts of whole bodies (Piñol et al., 2014). Hence, the DNA of interest are in minority, while the unexciting consumer DNA will compose a competitive majority.
Conceptually different approaches have been developed to enable prey studies from such “mixed” DNA samples. Nowadays most popular approach was spurred when Nielsen et al., (1991) researched synthetic analogs to DNA. A polymer with peptide instead of a sugar-phosphate backbone showed particular promise, because it formed stable hybrid duplexes with DNA (Nielsen et al., 1991). So-called peptide nucleic acids (PNAs) had higher melting temperatures than DNA (Egholm et al., 1993), went unrecognized by DNA polymerases, and could not initiate amplification by PCR (Orum et al., 1993). PNAs could be introduced prior to PCR to hybridize irreversibly with a target sequence and thereby suppress its amplification (Orum et al., 1993). Comparatively rare but interesting sequences would thus be allowed to replicate to detectable abundances (e.g. eukaryote parasites of blue crab, Troedsson et al., 2008). Other variants of blocking primers have also been put to the test, such as oligonucleotides modified with inhibitory C3 spacers (Deagle et al., 2009; Vestheim & Jarman, 2008). Since then, blocking primers have been used to study the prey of many different animals and with sample material ranging from whole body extracts of copepods (Cleary et al., 2016, 2017; Durbin & Casas, 2014; Novotny et al., 2021; Ray et al., 2016; Zamora-Terol et al., 2020), to dragonflies and apex canine predators (Morrill et al., 2021; Shi et al., 2021).
Although blocking primers have enabled many prey studies there are issues that warrant attention. Like universal PCR primers typically used in metabarcoding, blocking primers can introduce bias during amplification (Elbrecht & Leese, 2015; Leray & Knowlton, 2017; Piñol et al., 2015). With universal primers, primer-template mismatches and stochasticity may result in skewed relative abundances of species of interest (Sipos et al., 2007). Blocking primer bias also relates to primer-template mismatches, but it is rather the lack of them that leads to unreliable results. A blocking primer should have zero mismatches with the consumer, and as many as possible mismatches with prey to limit off-target interactions. Piñol et al., (2015) showed that a blocking primer with four and five mismatches to interesting prey decreased their relative abundances. Hence, PNAs and blocking oligonucleotides may introduce strong taxonomic biases during amplification (Piñol et al., 2015, 2019). Furthermore, production of specific and ultimately successful blocking primers remains an expensive chemical procedure, and relies on both consumer and prey sequences before the design can begin.
We have explored a simple, cost-efficient, and versatile approach to prey metabarcoding of small invertebrate consumers. By sequencing deep to offset for the overabundance of consumer DNA, we show that one can obtain ample prey reads from mixed whole-body copepod extracts, while at the same time avoiding the costs and laborious design of potentially biased blocking primers. Conceptually, we argue that the brute force method holds a lot of promise because current sequencing and computation enable acquiring and processing large amounts of data, and continuous development will only improve these capabilities (Lightbody et al., 2019). Through two sequencing runs (a pilot and a full-scale) using two commercially available NGS platforms (Illumina HiSeq4000 and Illumina NovaSeq6000), we tested the brute force approach for the first time with marine invertebrates. We report on the effect of sequencing depth for resolving prey composition and discuss advantages and caveats of the brute force methodology for prey studies of small consumers.