For all audiences: Incorporating immature stages into standardised inventories of mega-diverse groups has a major impact on our understanding of biodiversity patterns
Marc Domènech1, Owen S. Wangensteen2, Alba Enguídanos1, Jagoba Malumbres-Olarte3,4 & Miquel A. Arnedo1
1 Department of Evolutionary Biology, Ecology and Environmental Sciences & Biodiversity Research Institute (IRBio), Universitat de Barcelona, Av. Diagonal 643, E-08028, Barcelona, Spain.
2 Norwegian College of Fishery Science, University of Tromsø, The Arctic University of Norway, Norway.
3 cE3c – Centre for Ecology, Evolution and Environmental Changes / Azorean Biodiversity Group and Universidade dos Açores – Faculty of Agrarian and Environmental Sciences, Rua Capitão João d’Ávila, São Pedro, 9700-042 Angra do Heroísmo, Terceira, Açores, Portugal.
4 LIBRe – Laboratory for Integrative Biodiversity Research, Finnish Museum of Natural History, University of Helsinki, P.O.Box 17 (Pohjoinen Rautatiekatu 13), 00014 Helsinki, Finland
Correspondence : Marc Domènech, Department of Evolutionary Biology, Ecology and Environmental Sciences & IRBio, Universitat de Barcelona, Barcelona, Spain. Email: mdomenan@gmail.com
ABSTRACT
Because of their challenging taxonomy, arthropods are traditionally underrepresented in biological inventories and monitoring programs. However, arthropods are the largest component of biodiversity, and no assessment can be considered informative without including them. Arthropod immature stages are often discarded during sorting, despite frequently representing more than half of the collected individuals. To date, little effort has been devoted to characterising the impact of discarding non-adult specimens on our diversity estimates.
Here, we use a metabarcoding approach to analyse spiders from white oak communities in the Iberian Peninsula collected with standardised protocols, to assess (1) the contribution of juvenile stages to local diversity estimates, and (2) their effect on the diversity patterns inferred across communities. We further investigate the ability of metabarcoding to inform on abundance. We obtained 363 and 331 species as adults and juveniles, respectively. Species represented only by juveniles represented an increase of 35% with respect to those identified from adults in the whole sampling. Differences in composition between communities were greatly reduced when immature stages were taken into account, especially across latitudes. Moreover, our results revealed that metabarcoding data are to a certain extent quantitative, but some sort of taxonomic conversion factor may be necessary to provide accurate informative estimates.
Although our findings do not question the relevance of the information provided by adult-based inventories, they also reveal that juveniles provide a novel and relevant layer of knowledge that, especially in areas with marked seasonality, may influence our interpretations, providing more accurate information from standardised biological inventories.
Keywords : Araneae, Diversity, DNA barcoding, Iberian Peninsula, Metabarcoding, Spiders
Running title: Incorporating juveniles in diversity estimates
INTRODUCTION
Global human activity is altering the species richness and abundance of biological communities (Stuart Chapin III et al 2000; Socolar, Valderrama-Sandoval, & Wilcove, 2019; Barlow et al., 2016). These disturbances are accelerating the invasion rate of exotic species (Hulme, 2009) and driving numerous species to extinction, sometimes even before they are described, in what has been referred to as thesixth mass extinction (Dirzo et al, 2014; Barnosky et al, 2011). Bioinventories and early detection methods for monitoring ecosystem changes are essential to identify and tackle unanticipated threats to biodiversity (Telfer et al 2015; Barnosky et al 2013). However, sampling and identifying highly abundant and diverse groups such as arthropods is a daunting task. Several methods have been devised to overcome this limitation. For instance, rapid biodiversity assessment protocols are fast yet efficient sampling strategies specifically designed to retrieve the greatest amount of information from a particular area, minimising the number and length of sampling periods (Oliver and Beattie 1996).
Arthropods represent the largest and most abundant component of animal biodiversity. Therefore, no biodiversity monitoring program can be considered credible unless it takes them into account (Taylor and Doran 2001). Moreover, because of their high reproductive rates and short generation times, arthropods have the potential to inform on biodiversity changes at finer spatial and shorter temporal scales than vertebrates (Kremen et al 1993, Yen & Butcher 1997). However, the poor taxonomic knowledge of many arthropods limits their use as bioindicators, and they are frequently underrepresented in biodiversity assessments and conservation programs (Cardoso, Borges, Triantis, Ferrández, & Martín, 2012). On the other hand, recent studies suggest that arthropod populations, which play a fundamental role in ecosystem functioning, are declining at an alarming rate (Leather 2018).
Because of their rapid and divergent evolution, male copulatory apparatus and, to a lesser degree, female external reproductive systems are the main structures for species identification across most arthropod groups (Eberhard, 1985). Spiders are no exception. Morphological taxonomic identification is almost exclusively based on genitalic characters, i. e., the structure of the copulatory bulb in males and the vulva and epigyne (external modifications of the genital area) in females. These features are only visible in the last moult, which makes immatures difficult or impossible to identify at species level (Dobyns, 1997; Coddington, Young, & Coyle, 1996). Thus, in most inventories and diversity studies immature stages are discarded during sorting. However, immature specimens may account for between 40% and 70% of the collected specimens in biodiversity surveys (Soukainen et al, 2020, Malumbres-Olarte et al, 2020a, Malumbres-Olarte, Cardoso, & Crespo, 2019; Cardoso, Silva, Oliveira, & Serrano, 2004; Russell-Smith & Stork 1995, Silva 1996), or even up to 94% in extreme cases (Kuntner & Baxter, 1997). Disregarding immatures may significantly influence the inference of the temporal and spatial patterns of biodiversity, so their incorporation is desirable to obtain reliable estimates of diversity in short-term sampling protocols (Toti, Coyle, & Miller, 2000; Sorensen, Coddington, & Scharff, 2002).
It is known that the life cycles of different groups of spiders differ in the number of generations per year and in the time of the year when they are present as adults or juveniles (Aitchison, 1984; Nadal, Achitte-Schmutzler, Zanone, Gonzalez, & Avalos, 2018), even within the same species. For example, some wolf spiders are known to have annual life cycles maturing as adults in March-April and reproducing in May-June, although they can also have two clutches in the same year depending on the weather conditions (Rádai, Kiss, & Samu, 2017). Since rapid biodiversity assessment protocols are usually conducted once, they provide a “photograph” of the species present in a certain area at one particular time, so considering only adult individuals would completely dismiss all the species that are present as immature stages in that particular time of the year. To our knowledge, only one study (Norris, 1999) has partially addressed the effect in diversity estimates of incorporating juvenile spider stages. This study, which included only a few species that could reliably be morphologically identified at immature stages, already pointed out that numerous species were only found as juveniles and that relative abundances changed drastically when these immature stages were taken into account.
The use of DNA-based approaches for species identification, e.g., DNA barcoding (Hebert, Ratnasingham, & DeWaard, 2003), ease the identification of immature stages. DNA barcode sequences of immature individuals can be assigned to species through comparison to reference databases containing barcodes of adult-based morphologically identified species (Richard et al, 2010; Meiklejohn, Wallman, & Dowton, 2012). However, this technique still poses some drawbacks. For example, it requires manually extracting and amplifying each specimen individually, which for large samplings with hundreds or thousands of juveniles can be very time-consuming. Moreover, the economic cost of extracting, amplifying and sequencing such a large amount of samples would also be considerable.
DNA metabarcoding is a more recently developed molecular technique consisting in the automated identification of multiple species from a single bulk sample containing entire or partial organisms or from environmental samples (water, soil, etc.) containing remains of DNA (e.g., Bohmann et al., 2014; Yu et al., 2012; Morinière et al., 2016). This approach represents a clear advantage with respect to DNA barcoding, as it allows the simultaneous processing of a large number of specimens at once, greatly reducing the workload and processing time. In addition, it is more cost efficient for large numbers of specimens, as the number of sequences obtained from a single metabarcoding run is in the order of millions (Sales et al., 2020; Watts et al., 2019). The downside of using this approach is that individual specimens cannot be traced back or are sometimes even lost in the process of preparing the bulk sample, making it impossible to revise the voucher specimens if interesting sequences were found.
Another potential drawback of the use of metabarcoding for biodiversity assessment is its presumed inability to provide abundance information. To what extent the number of sequence reads of a certain taxon correctly represents its abundance or biomass in the sample has been a matter of much debate. Several studies have specifically addressed this issue (Elbrecht and Leese, 2015; Piñol, Mir, Gomez-Polo, & Agustí, 2015; Lamb et al., 2019; Deagle et al., 2019), but the answer remains inconclusive. While some studies consider the quantitative power of metabarcoding limited (Elbrecht and Leese, 2015; Piñol et al., 2015), others found metabarcoding to give an accurate estimate of a taxon’s abundance under certain conditions (Ratcliffe et al., 2020) or by applying correction factors that may vary among taxa (Kennedy et al., 2020; Thomas, Deagle, Eveson, Harsch, & Trites, 2016), or have even used it to quantitatively analyse dietary data (Soinien et al., 2015). One explanation for these different conclusions may be that the range of concentrations analysed varies considerably across studies, as suggested by Deagle et al. (2019). While a positive relationship between number of reads and biomass is commonly found, only a certain part of the variation in the number of reads seems to be explained by differences in the biomass in the sample, while the rest of the variation seems to be due to factors such as primer specificity (Elbrecht and Leese, 2015), different extraction success between different tissues or species (Schiebelhut, Abboud, Gómez-Daglio, Swift, & Dawson, 2016) or the efficiency of the blocking primers of predator DNA in the case of diet studies (Piñol et al., 2015).