Using DNA metabarcoding as a novel approach for analysis of platypus diet

Platypuses (Ornithorhynchus anatinus) forage for macroinvertebrate prey exclusively in freshwater habitats. Because food material in their faeces is well digested and mostly unidentifiable, previous dietary studies have relied on cheek pouch assessments and stable isotope analysis. Given DNA metabarcoding can identify species composition from only fragments of genetic material, we investigated its effectiveness in analysing the diet of platypuses, and to assess variation across seasons and sexes. Of the 18 orders and 60 families identified, Ephemeroptera and Diptera were the most prevalent orders, detected in 100% of samples, followed by Trichoptera, Pulmonata, and Odonata (86.21% of samples). Caenidae and Chironomidae were the most common families. Diptera had a high average DNA read, suggesting it is an important dietary component that may have been underestimated in previous studies. We found no variation in diet between sexes and only minimal changes between seasons. DNA metabarcoding proved to be a highly useful tool for assessing platypus diet, improving prey identification compared to cheek pouch analysis, which can underestimate soft-bodied organisms, and stable isotope analysis which cannot distinguish all taxa isotopically. This will be a useful tool for investigating how platypus prey diversity is impacted by habitat degradation as a result of anthropogenic stressors.

Platypuses prey on a wide variety of benthic macroinvertebrates such as insects, crustaceans, worms, and molluscs, with the orders of Trichoptera, Ephemeroptera, Odonata, and Coleoptera most commonly consumed 25,[27][28][29][30] . Their diet has been shown to vary seasonally in the wild 25,28 and captivity 31,32 , with no reported differences between males and females.
We used DNA metabarcoding for the first time on platypus check pouch samples, collected from the Snowy and Upper Murray Rivers regions, to determine its viability for assessing the species' diet. We predicted that this would be a successful approach in assessing platypus diet, while also potentially resulting in the detection of non-dominant orders and additional orders compared to those detected through morphological cheek pouch analysis. We also compared seasonal differences in diet between summer (breeding season, September-February) and autumn (non-breeding season, March-August) and between sexes. We predicted some variation between seasons and potentially between sexes in response to differences in metabolic demands between males and females and variable prey abundance throughout the year.

Sample collection.
Cheek pouch samples were collected from platypuses captured during surveys 33 using unweighted mesh nets or fyke nets. Mesh nets were set parallel to riverbanks in large, deep pools (> 50 m long, 1-2 m deep), from dusk until 1.00 A.M. and checked every few minutes with a spotlight in additional to a physical examination at least once an hour to remove possible snags. Platypuses were removed from the nets immediately upon being observed. Fyke nets were set in pairs in small shallow streams (< 1 m) in the late afternoon and checked every three hours until shortly after sunrise. Captured platypuses were transferred to pillowcases before processing, following established protocols 33,34 . Individuals were anesthetized in an induction chamber using isoflurane (Pharmachem, 5%) in oxygen (3 L/min), over 5-7 min 35,36 . Anaesthesia was maintained using a T-piece face mask, with isoflurane (1-1.5%) in oxygen (1.0L/min) 37 . Macroinvertebrate samples were collected from platypuses by inserting a small stainless-steel spoon into the buccal cavity via the bill 28  DNA sequencing and analysis. DNA was extracted from platypus cheek pouch samples (approx. 0.05 cm 3 of sample used after homogenisation) using the DNeasy Blood and Tissue Kit (QIAGEN, Hilden, Germany), following instructions from the manufacturer. All PCR and sequencing steps were performed by Mr DNA (Molecular Research, Shallowater, TX, USA; http:// www. mrdna lab. com). A 313 base pair (bp) region of the COI gene was amplified using the primers mlCOIintF (GGW ACW GGW TGA ACW GTW TAY CCY CC) and jgHCO2198 (TAIACYTCIGGRTGICCR AAR AAYCA) 38,39 . Unique index barcodes of 8 bp were attached to each primer. One-step PCRs were done in duplicate, using the HotStarTaq Plus Master Mix Kit (Qiagen, USA). Thermocycling protocols consisted of 94 °C for 3 min; 30 cycles of 94 °C for 30 s, 53 °C for 40 s and 72 °C for 1 min; and a final elongation step at 72 °C for 5 min. Successful amplification was checked on a 2% agarose gel. Samples were pooled in equal proportions, based on molecular weight, purified using Ampure XP beads and an Illumina DNA library prepared. Samples were sequenced on an Illumina MiSeq sequencer, using V2 300 cycle kit and PhiX spike-in was added at 20%. Taxonomic assignment. Sequence data from each sample was demultiplexed using custom-built FASTq Processor script (http:// www. mrdna frees oftwa re. com/). The Greenfield Hybrid Analysis Pipeline (GHAP) 40 was used for processing data and creating operational taxonomic units (OTU) tables. Sequences with a minimum overlap of 25 bp and homology of at least 80% were merged and only sequences between 301 and 313 bp (i.e. close to the target amplicon length) were retained for further analyses 41 . Sequences were clustered into OTUs, using a 97% clustering threshold as this threshold corresponds reasonably well with COI delimitation between invertebrate species 42 . To remove possible sequencing artifacts, OTUs that made up less than 0.01% of the total reads in a sample were filtered out.
OTUs were associated with taxa by taking the highest percent homology matches from a reference library of global sequences from GenBank 43 (https:// www. ncbi. nlm. nih. gov/ genba nk/, accessed 10th of June 2021). We then refined this taxonomic assignment using percent homology thresholds, arbitrarily assigned as part of the GHAP pipeline: 97% or greater for species, 95 to < 97% for genus, 90 to < 95% for family, 85-< 90% for order, and those with < 85% were not assigned because of taxonomic uncertainty 41 .
Freshwater snails (families Ancylini, Agriolimacidae, Lymnaeidae, and Physidae) were all assigned to the informal group (previously an order) Pulmonata. This informal group has been included as an order in the analysis for ease of interpretation.
Analyses. To evaluate variation in diet between samples, seasons, and sex, we quantified four metrics for each taxon (order or family, depending on the taxonomic scale of analysis): (1) taxon diversity (average number of taxa in each sample); (2) taxon prevalence (the percentage of samples that each taxon was detected in); (3) taxon DNA reads (the average DNA read numbers of each taxon across all samples); (4) taxon relative preva- www.nature.com/scientificreports/ lence (the proportion of samples that each taxon was detected in, relative to the total number of taxa detected for that pool of samples (total, season or sex)). Due to the low number of samples collected during spring (n = 3) and winter (n = 2), we grouped samples from spring with summer, representing the breeding season (September-February) and samples from winter with autumn, representing the non-breeding season (March-August). Given the small sample sizes, and the nature of count and proportional data, we compared differences in the number of orders and families present in each sample between seasons and between sexes using non-parametric Wilcoxon Signed-Rank tests 44 . We also compared differences in individual taxon DNA reads between seasons and sexes using Wilcoxon Signed-Rank tests. We plotted taxon accumulative curves using the 'rarefy' function in the 'vegan' package (v2.5-7) 45 . All analyses were implemented within R (v4.0.2) 46 .

Results
Metabarcoding effectiveness. A total of 2,346,944 reads were generated from the MiSeq run. Prior to analysis, several kingdoms, phyla, and orders were categorised as environmental DNA and removed from the analysis (Appendix 2). After data filtering, there were 1,389,491 reads attributable to invertebrates.

Seasonal differences.
A total of 18 orders and 53 families were detected in samples collected in summer (September-February), compared to 15 orders and 43 families detected from samples collected in autumn (March-August). The average number of orders in the samples was not significantly different between summer (8.19 ± 0.60 se) and autumn (9.38 ± 0.31 se, W = 136.5, P = 0.154). In Summer, orders Diptera and Ephemeroptera were detected in 100% of samples, while in autumn, Diptera, Ephemeroptera, and Trichoptera were all detected in 100% of samples (Fig. 5, Appendix 4). Diptera had the highest average DNA read for summer (14,121 ± 4727) and Ephemeroptera had the highest average DNA read for autumn (10,007 ± 3114 se). DNA reads only differed significantly between seasons for Diptera, higher in summer (W = 58, P = 0.044; Fig. 5) and Plecoptera (stoneflies), higher in autumn (W = 163, P = 0.003). The largest difference in the relative prevalence of orders between seasons was for Plecoptera more prevalent in autumn (7.37% of diet) compared to summer (5.53% of diet; Fig. 4) and Haplotaxida (9.92% of diet in summer, 4.10% in autumn). Orders Amphipoda, Decapoda, and Hirudinida were only detected in summer, although in a small number of samples (Fig. 5).
There was no difference in the number of families detected in samples between summer (11.63 ± 1.59 se) and autumn (13.08 ± 1.18 se, W = 132, P = 0.226). Families Caenidae and Chironomidae were most prevalent during summer (detected in 100% and 87.5% of samples, respectively), while during autumn Caenidae (100%) and Gomphidae (84.62%) were most prevalent. Caenidae had the highest average DNA read within the samples during summer (11,702 ± 5267 se) and Baetidae (mayflies) had the highest for autumn (5780 ± 2363 se). Average DNA reads were significantly higher in summer for Chironomidae (W = 52, P = 0.023). The largest differences in relative prevalence of families between seasons was for Gripopterygidae (stoneflies, 1.08% of diet in summer and 4.71% in autumn).
Sex differences. Eighteen orders and 53 families were detected in the diet of male platypuses, compared to 15 orders and 44 families in the diet of females. The number of orders detected in the diet of males (8.72 ± 0.51 se) was not significantly different than females (8.73 ± 0.54 se; W = 96, P = 0.909). Diptera and Ephemeroptera were the most prevalent for both sexes, detected in 100% of samples (Appendix 5). Ephemeroptera had the highest average DNA read across samples for both females (15,525 ± 6651 se) and males (9134 ± 3113 se). There was no significant difference in the average DNA reads for any order between males and females (Fig. 5). The relative prevalence of orders between sexes was similar for most orders (Fig. 4). www.nature.com/scientificreports/ There was also no difference in the number of families between males (12.72 ± 1.46 se) and females (11.54 ± 1.25 se; W = 99, P = 1.00). Caenidae was the most prevalent family in the diet of both males and females (detected in 100% of samples) and also had the highest average DNA read for females (10,924 ± 6463 se) and males (6527 ± 3247 se). There was no significant difference in the DNA read for any family between males and females (Fig. 5). The relative prevalence of families between sexes was similar for most orders.

Discussion
This study is the first to use DNA metabarcoding to assess the diet of platypuses, resulting in the detection of 18 orders and 60 families. The most common orders in the diet of the 29 platypus cheek pouch samples we analysed were Ephemeroptera and Diptera, and the most common families were Caenidae and Chironomidae. The number of orders and families in each sample was not significantly affected by season or sex. As in other studies using morphology-based cheek pouch identification and stable isotope analysis of platypus diet (Table 1), the orders Ephemeroptera, Diptera, Trichoptera, and Odonata were all important components of platypus diet, in terms of prevalence and average DNA read 25,[27][28][29] (Fig. 3).
Previous cheek pouch studies using morphological identification indicate that Diptera were common in the diet of platypuses, but rarely dominant (Table 1), with the conclusion that they are likely avoided because of their small size or inaccessibility 28,29 . Stable isotope and our DNA analyses indicate that they were a dominant component of platypus diet, present in 100% of samples and having the second highest average DNA read across www.nature.com/scientificreports/ samples (Fig. 3, Appendix 3). This suggests that they may have been underestimated in studies using morphologybased cheek pouch identification, possibly due to increased mastication of these soft-bodied organisms 27,30 . The presence of Diptera in pools where platypuses preferentially forage suggests that they are likely one of the most important dietary components for platypuses 27 . This is supported by the lack of worms in the platypus diet reported from previous studies, despite Marchant and Grant 30 noting their high abundance in the benthos. Isotope analyses suggested that the combination of Diptera and Odonata were a large dietary component, supporting their importance in the diet, but given they could not be separated isotopically, further investigations, such as our own, are required to establish the importance of each individual order. Coleoptera was also identified as an important component of platypus diet from other studies (Table 1), but they did not rank as highly from the DNA analyses, being the 6th most prevalent order. Interestingly, the order Pulmonata (snails), equal 3rd most prevalent order, was not identified in previous studies as a common order (Table 1). Previous assessments using morphology-based cheek pouch identification note the difficulties in comparing the finely ground shells of bivalves and gastropods 29 , potentially leading to underestimation in previous studies, with DNA analysis also revealing the order Veneroida (bivalves) were a relatively important dietary component. Reported differences between orders and families in the diet of platypuses between studies may reflect differences in local habitats and macroinvertebrate availability, but may also highlight differences resulting from identification methodology.
While the number of distinct orders and families was higher in summer compared to autumn, there was no significant difference in the average number of orders/families in the samples between these two seasons. Previously reported seasonal differences have been between summer/autumn and winter/spring, when differences in metabolic requirements and prey availability may be greater 25,28 , possibly explaining why we found no differences in the diet when comparing summer and autumn. Only Diptera and Plecoptera had significant differences in the average DNA reads between seasons, possibly reflecting availability.
There have been no reported differences in the diet of male and female platypuses 25,27,28 . DNA analysis suggests males and females had the same dominant orders of Ephemeroptera and Diptera in their diet. McLachlan-Troup et al. 25 suggest that given males are larger than females, there may be an expected difference in number of taxa consumed between sexes, but that this may not have been detected due to the mastication of prey items in the cheek pouches. However, stable isotope analysis and this DNA analysis also confirms no differences in diet between sexes.
This study was limited by a small sample size and because the samples used in this study were obtained from a number of rivers across two regions (Fig. 1), which may have confounded results given local differences in macroinvertebrate assemblages. Pooling samples into breeding (spring/summer) and non-breeding (winter/ autumn) breeding seasons may have also confounded seasonal differences in diet. Future studies can provide greater clarity on dietary selection across seasons by undertaking macroinvertebrate surveys in conjunction www.nature.com/scientificreports/ with cheek pouch analyses. We also did not undertake morphological analyses of the cheek pouch samples to make comparisons with DNA metabarcoding results. However, given the high sensitivity of this method, the high number of reported orders and families from a relatively small sample size, and similar reported taxa to previous studies, we expect this method would have resulted in the detection of more orders and families than morphological analysis of the cheek pouch samples. The spoon used to collect cheek pouch material was sterilised using 100% ethanol between samples, but was not flame sterilised, which may have contributed to some contamination between samples. DNA metabarcoding is clearly a useful tool for assessing the diet of platypuses. It improves the ease of identification compared to morphology-based cheek pouch analysis and may increase detection of soft-bodied organisms which may be underestimated using morphology-based methods. While isotopic analysis is also an effective method for assessing platypus diet, variation can occur depending on the part of body sampled, and taxa cannot always be separated isotopically 27 . Future studies should aim to maximise sample sizes to allow for more accurate comparisons between seasons and sexes. There are further opportunities for using DNA metabarcoding for assessing platypus diet. It could be used effectively to compare differences in diet across regions and between rivers, as well as for assessing platypus diet in degraded habitats or downstream of heavily regulated rivers where platypuses are suspected to be declining due to changes in the macroinvertebrate composition 33 .  License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.