Research Article |
Corresponding author: Stephanie J. Swenson ( stephanie.swenson@uni-kassel.de ) Corresponding author: Birgit Gemeinholzer ( b.gemeinholzer@uni-kassel.de ) Academic editor: Chloe Robinson
© 2022 Stephanie J. Swenson, Lisa Eichler, Thomas Hörren, Andreas Kolter, Sebastian Köthe, Gerlind U. C. Lehmann, Gotthard Meinel, Roland Mühlethaler, Martin Sorg, Birgit Gemeinholzer.
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Citation:
Swenson SJ, Eichler L, Hörren T, Kolter A, Köthe S, Lehmann GUC, Meinel G, Mühlethaler R, Sorg M, Gemeinholzer B (2022) The potential of metabarcoding plant components of Malaise trap samples to enhance knowledge of plant-insect interactions. Metabarcoding and Metagenomics 6: e85213. https://doi.org/10.3897/mbmg.6.85213
|
The worldwide rapid declines in insect and plant abundance and diversity that have occurred in the past decades have gained public attention and demand for political actions to counteract these declines are growing. Rapid large-scale biomonitoring can aid in observing these changes and provide information for decisions for land management and species protection. Malaise traps have long been used for insect sampling and when insects are captured in these traps, they carry traces of plants they have visited on the body surface or as digested food material in the gut contents. Metabarcoding offers a promising method for identifying these plant traces, providing insight into the plants with which insects are directly interacting at a given time. To test the efficacy of DNA metabarcoding with these sample types, 79 samples from 21 sites across Germany were analysed with the ITS2 barcode. This study, to our knowledge, is the first examination of metabarcoding plant DNA traces from Malaise trap samples. Here, we report on the feasibility of sequencing these sample types, analysis of the resulting taxa, the usage of cultivated plants by insects near nature conservancy areas and the detection of rare and neophyte species. Due to the frequency of contamination and false positive reads, isolation and PCR negative controls should be used in every reaction. Metabarcoding has advantages in efficiency and resolution over microscopic identification of pollen and is the only possible identification method for the other plant traces from Malaise traps and could provide a broad utility for future studies of plant-insect interactions.
biomonitoring, DNA metabarcoding, insect decline, landscape change, nature conservation, plant-insect interactions
Landscape level change and chemical input in agriculture are major contributors to the rapid level of decline in diversity and abundance of insects observed in recent decades (
Due to these declines and the fast pace of landscape changes, development of novel methods of monitoring flora in order to understand which plant resources are directly used by insects on a temporal scale, could lead to better pest management practices in agriculture, as well as land management decisions for scale and spacing of conservation areas. Malaise traps have been used for collection of flying insects for more than 80 years (
In spite of this potential, traditional microscopy identification of the pollen contained in Malaise traps is extremely time-consuming and requires extensive training and identification of plant fragments or regurgitated food material is nearly impossible. However, advancing techniques in genetic identification of complex mixed species environmental samples could provide a potential resource to identify the plant components found in Malaise trap samples. Metabarcoding, using a short gene region, or barcode, for the identification of many taxa contained in a complex sample, has displayed a great potential for plant and pollen identification over the last decade (
In addition to the challenges inherent in plant barcoding and metabarcoding, the sample type presented by Malaise traps differs from those of previous studies in that the sample contains two signal types: 1) pollen and plant material as well as eDNA carried in on the insect body and 2) partially digested plant and pollen material excreted from the digestive tract or released due to breakage after capture. It is uncertain how these signals may interfere with each other and bias results.
This study aims to address whether Illumina MiSeq metabarcoding of plant fragments and eDNA found in Malaise trap preservative ethanol using the ITS2 barcode can retrieve a realistic assemblage of the vegetation available and utilised by the insects found in the traps. In addition, we examine whether rare or neophyte species can be detected in the samples and whether crop and non-native garden ornamentals are found in traps internal to protected areas indicating travel out of protected area for foraging. To evaluate our results, we developed the following hypotheses: 1) taxa retrieved will represent a realistic assemblage of German native plant taxa and complement the vegetation surveys taken in the sampling area; 2) plant species retrieved will primarily be those with pollen available in the respective regions and sampling duration; 3) insects will travel into and out of protected areas to forage and evidence of non-native garden plant and crop plant species will be found in the internal-most trap in the nature conservancy areas and 4) threatened and endangered plant species will be detected in the samples, but likely in very low read quantities.
As part of the project DINA (Diversity of Insects in Nature Protected Areas) (
Map of the Malaise trap sampling sites throughout Germany, with site codes and names. Green indicates sites with complete data for all five traps, blue sites are missing data from at least one trap and orange indicates no data for the site.
At each of the 21 sites, five Malaise traps were placed at a gradient with the first located 25 m into arable land or as close as possible when the landscape would not permit. Subsequent traps were located 25 m distance from the other, with the second trap located directly on the intersection of arable and protected land and the fifth located 75 m towards the centre of the nature protected area (
Insects were collected using the standardised sampling design of German long-term studies of insect biomass (
The original sample ethanol (200 ml ± 50 ml) was vacuum filtered using a 250 ml Nalgene single use analytical filter funnel with a cellulose nitrate (CN) filter (diameter 47 mm and 0.2 µl pore size) in a biosafety cabinet with DNA free equipment. Following filtration, the CN filter was cut into two equal parts and each part placed in a 2 ml SafeSeal microcentrifuge tube (Sarstedt AG & Co. KG), with one half used for DNA extraction and the other saved as a voucher and/or backup for protocol optimisation. These samples were stored at -20 °C until further processing.
DNA extraction was performed with NucleoMag 96 Plant Kit (Macherey Nagel, Oesingen, Switzerland) with the following changes to the standard protocol: 1) 1 gm of 1.4 mm ceramic beads, 500 µl lysis buffer MC1, 5 µl Proteinase K (Macherey Nagel, Oesingen, Switzerland), 5 µl RNaseA were added to the 2 ml microcentrifuge tube containing the half filter paper with sediment and tissue was disrupted for 2.5 minutes with a Retsch MM400 bead mill at 30 Hz.; 2) following homogenisation, samples were incubated at 65 °C for one hour with constant shaking in addition to manual inversion mixing of the tubes every ten minutes to ensure uniformity of sample lysis; 3) following incubation, samples were centrifuged for ten minutes and the resulting 250–300 µl of lysate were transferred to clean 2 ml tubes; 4) 300 µl of binding buffer MC2 and 15 µl of magnetic beads were added; 5) remaining reagents were used at 25% of the standard protocol, with the exception of the elution buffer MC6 of which 35 µl were added and incubated at 50 °C for five minutes to evaporate any residual ethanol then 6) 25 µl were removed for PCR and sequencing and 2 µl for DNA quantification with Qubit 4 fluorometer (Thermo Fisher Scientific Inc.).
The ITS2 barcode was chosen for its high rate of success for species level identification, as well as having amongst the most abundant reference sequences available in public DNA sequencing repositories (
Sequencing data were processed with USEARCH (
The database was created in June 2021 from sequences downloaded using the GenBank webinterface at https://www.ncbi.nlm.nih.gov/genbank/ in GenBank (full) format using the search string (internal transcribed spacer1[Title] OR internal transcribed spacer 2[Title] OR ITS1[Title] OR ITS2[Title]) NOT patent NOT pseudogene NOT mRNA NOT unverified AND 100:2500[Sequence Length] AND Tracheophyta[Organism]. The term unverified is used by GenBank staff to flag erroneous sequences and should be excluded in every GenBank query. Due to the size of the downloaded data file, it was loaded into R by using the function fread from the package data.table. Taxonomic classifications and sequence information were extracted per GenBank accession number in the database. Species names were cleaned by removing any subspecies or variety information. Subsequently, unclassified environmental sequences and sequences containing any of the following characters [!§$%&/()=?`*’+#;`]., except for hyphen and underscore (the internally used string separator), were removed. Other irregularities which were removed include the presence of the keyword Eukaryota at the family descriptor and names starting with x_ or ending with a number. All plant family names were compared to the taxonomic backbone of Global Biodiversity Information Facility (GBIF) and discarded if no match with the status keyword ACCEPTED were found. The subsequent step removed all sequences with more than 1% ambiguous nucleotides. The ITSx algorithm was set to determine the stop and start positions of ITS1 and ITS2 sequences which matched a predefined Tracheophyte HMMER profile, also supplied by ITSx (
ASVs with ambiguous species level identifications were given only genus level identifications. Taxa that do not occur in Germany and were likely the result of laboratory contamination were removed from analyses. Fungal contaminants were confirmed with BLAST search and removed. Values of Amplicon Sequence Variants (ASVs) found in extraction and PCR blanks were used to establish a relative abundance per negative threshold and ASVs not occurring above the highest relative abundance were removed. ASVs with less than five reads were removed.
DNA yield of all samples (105) tended to be low with a range of 0.152 to 13.700 ng/µl (mean = 1.41, s.d. = 1.69). The majority of samples with a DNA yield below 0.5 ng/µl failed to amplify or produced low read abundances and were removed from analysis. Of the 21 sites, 10 sites produced full sequencing data for all five Malaise trap samples, 10 sites had 1–4 traps fail to amplify and one site had all traps failing to amplify. Attempts to amplify failing samples with different conditions have not yet been successful. Overall, 78 of the 105 samples provided sequencing data. The four species mock community positive controls retrieved all species and these species were absent in the Malaise trap samples. Extraction blanks and PCR blanks indicate a very low level of cross contamination. Plant reads were only rarely present (1, 3 and 90 reads in the three replications) in PCR blanks. Plant reads were more common in extraction blanks; however, they were limited to the most common taxa in a low read number (≤ 30) or were taxa only found in the extraction blanks (Fraxinus excelsior and Cucumis sativus).
We identified 60 plant families in our reads, with 223 genus level identifications and 243 species level identifications (Suppl. material
The vegetation surveys recovered 48 flowering plant families, three of which were not recovered from metabarcoding (Orchidaceae, Hypericaceae and Linaceae). When extended to genera present, 77 were recovered only from vegetation surveys, 104 were recovered only from metabarcoding and 119 were recovered from both metabarcoding and vegetation surveys (Suppl. material
Species level identification of Brassica spp. ASVs was not possible due to their hybridogenous and polyploid origin. Based on the potential presence of B. napus, B. nigra, B. oleracea, and B. rapa in the sampling area and the hybridogenous DNA of Brassica cultivars, all of these taxa are likely included in the Brassica spp. ASV umbrella. Due to the prevalence in agriculture throughout Germany and May flowering time, we expect B. napus to be the most abundant species in the areas surrounding our experimental sites.
The highest generic level diversity was displayed in Brassicaceae (29), Poaceae (27), Asteraceae (21), Fabaceae (17), Rosaceae (12) and Caryophyllaceae (10). Of the taxa that could be assigned to at least generic level, only six were not likely to not have pollen available in the sampling time, the late summer or early autumn flowering Hedera helix and Helianthus annus and the late winter or early spring flowering Alnus spp., Carpinus betulus, Corylus spp. and Taxus spp.
We retrieved data for 14 Malaise traps placed most internal to the nature conservation area of the 21 sites. In these traps, we detected 21 agricultural or garden plants (Table
Garden and Agricultural plants found in Malaise traps located most internal to nature conservancy areas with numbers indicate percentage of read abundance of the species in the sample.
Taxonomy | Sites | |||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Family | Genus | Species | BOT | DOE | GEE | GIP | HOF | IPF | KOO | KOP | KUE | MAL | MIT | MUE | RIE | SCH |
Garden plants | ||||||||||||||||
Adoxaceae | Viburnum | opulus | 0 | 0 | 0 | 0 | 0 | 0.04 | 0 | 0 | 0.03 | 0 | 0 | 0 | 0 | 0 |
Amaryllidaceae | Allium | ursinum | 0 | 0 | 0 | 0.43 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
Apiaceae | Chaerophyllum | roseum | 0 | 0 | 0 | 0 | 0 | 0.02 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
Apiaceae | Heracleum | dissectum | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0.05 | <0.01 | 0 | 0 |
Asteraceae | Achillea | biebersteinii | 0 | 0.16 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
Asteraceae | Helianthus | annuus | 0 | 0 | 0 | 0 | 0.04 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
Asteraceae | Pilosella | castellana | 0.53 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 5.39 | 0 | 0 | 0.30 | 0 | 0 |
Asteraceae | Symphyotrichum | cordifolium | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
Brassicaceae | Aubrieta | olympica | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0.17 | 0 | 0 |
Brassicaceae | Aubrieta | sp. | 0.01 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0.70 | 0 | 0 |
Brassicaceae | Aurinia | saxatilis | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0.01 |
Caryophyllaceae | Cerastium | alpinum | 0 | 0 | 0 | 0 | <0.01 | 0 | 0 | 0 | 0.12 | <0.01 | 0 | 0 | 0 | <0.01 |
Cyperaceae | Cyperus | diandrus | 0 | 0.06 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0.02 | 0 | 0 |
Ericaceae | Erica | arborea | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0.14 | 0 | 0 | 0 |
Fabaceae | Wisteria | sp. | 0 | 0 | 0 | 0 | 0 | 0.10 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
Oleaceae | Syringa | vulgaris | 0 | 0 | 0.02 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | <0.01 | 0 | 0 |
Poaceae | Poa | trivialis | 0 | 0.03 | 0 | <0.01 | 0.02 | 0.14 | 0 | 0.13 | 0.17 | 0.03 | 4.87 | 0.08 | <0.01 | 0 |
Solanaceae | Solanum | lycopersicum | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | <0.01 | 0 | 0 | <0.01 | 0 | 0 |
Agricultural plants | ||||||||||||||||
Brassicaceae | Brassica | sp. | 96.26 | 18.45 | 36.36 | 9.01 | 2.16 | 0.53 | 71.03 | 0 | 27.22 | 2.31 | 0 | 0.05 | 99.33 | 4.79 |
Poaceae | Secale | cereale | 0.04 | 0.10 | 0.03 | 0 | 0.03 | 0.24 | 0 | <0.01 | 0 | 2.41 | 0 | 0 | <0.01 | 0.02 |
Poaceae | Triticum | sp. | <0.01 | 0.20 | 0 | 0 | 0.05 | 0.02 | 0 | 0 | <0.01 | 0 | 0 | 0.02 | 0 | 0.02 |
Red List and neophyte taxa detection
We detected 22 species listed as threatened (
Red list and neophyte plant species detected from ITS2 metabarcoding reads of Malaise trap plant components. *Indicates sites where the quantity of reads was low (≤10).
Status | Species | Number of sites | Site locations |
---|---|---|---|
Red list V | Aira coaryphyllea | 1 | DOE |
Red list 3 | Alyssum alyssoides | 1 | SCH |
Red list V | Camelina microcarpa | 1 | GEE |
Red list V | Camelina sativa | 1 | GEE |
Red list 2 | Chenopodiastrum murale | 1 | POR* |
Red list V | Cynoglossum officinale | 3 | GEE, POR, WIP |
Red list V | Eleocharis unglumis | 1 | KOO* |
Red list V | Genista sagittalis | 2 | DOE, WIP |
Red list V | Helictochloasp. | 2 | BOT, POR |
Red list V | Hippocrepis comosa | 5 | HOF, IPF, KUE, MUE, WIP |
Red list V | Hottania palustris | 1 | LUE |
Red list V | Lotus tenuis | 4 | HOF, IPF*, KOP, MUE* |
Red list 3 | Melamprum arvense | 1 | GEE |
Red list 3 | Myrica gale | 1 | LUE |
Red list 3 | Onobrychis viciifolia | 6 | HOF, MAL*, MIT*, MUE, SCH, WIP* |
Red list 2 | Papaver hybrium | 1 | MAL |
Red list V | Primula veris | 3 | GEE, HOF*, IPF* |
Red list V | Ranunculus polyanthemos | GIP, KOP*, LUE* | |
Red list 2 | Saponaria ocymoides | 1 | DOE* |
Red list V | Scleranthis perennis | 1 | DOE* |
Red list 3 | Silene conica | 1 | MAL |
Red list 3 | Silene otites | 1 | SCH |
Neophyte | Anacyclus clavatus | 1 | BOT |
Neophyte | Campanula portenschlagiana | 1 | IPF |
Neophyte | Caragana sp. | 1 | SCH* |
Neophyte | Hesperis sp. | 2 | MUE, SCH |
Neophyte | Holcus annuus | 17 | BIS*, BOT*, DOE, GIP, HOF*, IPF, KOO*, KOP, KUE*, LUE*, MAL, MIT, MUE, POR*, RIE, WIP*, WIS |
Neophyte | Lolium persicum | 1 | WIS |
Neophyte | Medicago sativa | 6 | DOE*, KUE*, MAL*, MIT*, SCH, WIS |
Neophyte | Monarda didyma | 1 | MAL |
Neophyte | Pimpinella peregrina | 1 | BRA |
Neophyte | Poa infirma | 3 | LUE*, MAL*, RIE |
Neophyte | Robinia pseudoacacia | 9 | BIS, BOT, BRA, DOE, IPF, KOP, MAL, MIT, POR, |
Neophyte | Trifolium incarnatum | 2 | MAL, SCH |
Neophyte | Vicia pannonica | 1 | SCH |
The low number of reads of taxa in extraction and PCR negative controls that also appear in the Malaise trap samples indicate a very low level of contamination between samples and soundness of our laboratory protocols. Additionally, the presence of all four species, contained in the positive control samples while not appearing in the Malaise trap samples, adds evidence of sound methods of extraction, primer choice and PCR protocols. Nevertheless 27 of our 105 samples failed to amplify, were primarily non-target fungal species or produced a low quantity read abundance. There are several possible reasons for these failures, including low quantity DNA extraction yield, impurities in the trap ethanol, variations in ethanol concentration, storage time until filtration and mechanical interference from non-plant debris found in the sample (
Our samples recovered 223 genus level identifications and 243 species level identifications from 60 families. Diversity of species in our study was reflective of the species diversity within German taxa. Families with high species diversity like Asteraceae, Brassicaceae, Caryophyllaceae, Fabaceae, Poaceae and Rosaceae retrieved the largest number of species, while those with lower diversity retrieved (Suppl. material
The differences between the vegetation surveys and metabarcoding results are more pronounced when comparing at the generic level, where the number of genera recovered from both were 119, metabarcoding only 104 and vegetation surveys only 77. Ten tree genera were found only in metabarcoding and are likely to occur outside of the 50 m perimeter of the vegetation survey area.
The low diversity or absence of some taxa known to be in the area of the study sites can be attributed to several possible reasons. The first contributing factor is that Malaise traps are known to capture the largest proportions of the pollinator rich orders, Hymenoptera and Diptera (
There are several other possible reasons for the low diversity or absence of certain taxa in our study sites beyond pollen availability. The first being the plants present in a sample represent insect preference and site-specific availability. While there are hundreds of species we did not recover from metabarcoding that may be plentiful in Germany as a whole, they might not be present in our study site or within the insect foraging range of our sites. In addition, some sites had already been managed with sheep grazing prior to this sampling duration and the Deutscher Wetterdienst (DWD) reported May 2020 to be very dry, both of these factors contributing to reduced plant resources in our study sites. After consideration of the time and space availability of particular plant species, recovery from metabarcoding is still dependent on the optimisation of DNA extraction, as well as its ability to overcome PCR bias. While we implemented good practice protocols of multiple PCR replicates, inclusion of positive controls and use of plant specific PCR primers that have been tested for optimal species recovery for this region, PCR bias will still have a role in the abundance and presence or absence of certain taxa within a sample (
Our utilisation of the ITS2 barcode could also partially explain the low level of retrieval of plants without available pollen. The plant DNA available from herbivory of vegetative plant structures will have been degraded by the digestion process and may require a barcode of shorter length to be amplified. Sample types composed of degraded DNA, such as herbal and food products, permafrost samples, faeces and ancient DNA, are more successfully resolved when a short DNA fragment, such as the P6 loop of trnL is implemented (
Over half of our ASVs could not be identified to species level. Although the database used provided a high coverage for German taxa, estimated at 90%, based on growth and coverage of a previously used database for the German state of Bavaria created in 2014 (
Garden and agricultural plants
We detected several species of known garden plants and agricultural plants in the Malaise traps located most internally in the nature protected areas, indicating travel by a proportion of insects into urban and agricultural landscapes for foraging. Our results indicate a much larger foraging area than plants in the immediate vicinity of the Malaise traps, which is strengthened by the non-native ornamental garden plants present in our samples. Our study sites differ in their geographic vicinity to settlements and we do not have specific information on the plants grown in these areas, but it can be assumed they will offer a low resource for foraging compared to native plants in the area of the Malaise traps and the agricultural crops planted adjacent to them. The garden species represented in these samples are present in low read abundance and do not appear to be relevant to insect food availability in conservation areas at this time of year, but could be indicative of insect flight distances at times of lower pollen availability.
We also detected agricultural species, primarily Brassica spp., in the internal-most traps; however, the trap most internal in the nature protected area is 75 m from the edge of cultivated land, a distance that is within the recorded foraging range of most hymenopteran and dipteran pollinators (
The most prevalent agricultural group of plants in our samples were the Brassica spp. complex. The high level of Brassica spp. in the majority of samples confound data interpretation. The first problem being the complication in species level identifications from genetic methods due to the hybridogenous and polyploid origin within the genera. This limits the ability to use vegetation surveys to confirm real presence or absence with possible contamination from the field or laboratory processing or over-representation due to PCR bias. However, due to these species’ (B. napus, B. nigra, B. oleracea, and B. rapa) prevalence in cultivation throughout Germany, as well as pollen of B. napus being an attractive and protein rich (
The other two agricultural species found in traps were Poaceae species, Secale cereale and Triticum spp. and, while they were present in the majority of traps, they were generally present with very low read numbers. Unlike the Brassica species, it seems more likely that the presence of these species was not the result of insect foraging. Poaceae is an anemophilic family and the Stiftung Deutscher Polleninformationsdienst (www.pollenstiftung.de) recorded this family to be in high pollen flight during the collection duration. The presence of these agricultural taxa in the samples, as well as other Poaceae, may be due to the ubiquitous occurrence of pollen in the environment where insects passively pick it up on their bodies. Some presence can also be explained by accidental, but unavoidable, contamination by practitioners in the field and laboratory. We detected several other taxa in our samples that were in peak pollen flight over our sampling duration, Pinaceae (6.5%), Poaceae (8.8%), Rumex spp. (2.3%). Sambucus spp. (4%), Quercus spp. (0.6%) and Aesculus spp. (0.7%). These results indicate a need for careful interpretation of results and incorporation of blank samples, as presence of a species recovered from metabarcoding might not represent a purposeful act by an insect and/or could be an accidental introduction from processing steps.
Red List species
While we did detect several threatened and near threatened Red List species (
Neophyte detection
We detected 13 neophyte species in our samples, with generally higher read quantity and across more sites than the Red List species. The same caution in data interpretation and preliminary experimentation recommended for Red List species should be applied to detection of these taxa, especially as we do not know if the presence of these species in our samples occurred from plants present within or outside of the nature protected areas. Nevertheless, this result strengthens the potential of metabarcoding as a long-term biomonitoring tool for tracking and preventing the spread of deleterious plant species.
Overall plant taxa recovery
The results of metabarcoding for the most part were in line with the vegetation surveys when viewed from the family level and the three families not recovered from metabarcoding are likely due to flowering time or pollination strategy. Metabarcoding recovered 15 more families than found in the vegetation surveys, indicating plants that are not only found in the immediate sampling vicinity, but the greater insect foraging range. The dissimilarity of taxa is more pronounced when viewed from a generic level and, while some of this can be explained by the limitations of metabarcoding, it does not reach a level of failure of the methods. The majority of this dissimilarity in genera recovered can be explained by site specific availability, insect foraging range and flowering phenology. These results illustrate the complimentary nature of incorporation of metabarcoding into Malaise trap biomonitoring programmes, vegetation surveys can give the entire view of plants in the area, while metabarcoding could potentially enhance knowledge of what plants are being used at a given time and how these assemblages change over a growing season.
Our study illustrates the potential of Malaise trap plant metabarcoding as an additional tool for large-scale plant biomonitoring; however, cautious consideration of its limitations must be included in project design, data analysis and interpretation. Further experimentation must be undertaken to account for sample failure with mock communities and examination of several barcodes and primer combinations should be evaluated. When specific species are of great interest to the particular study, preliminary experimentation with mock communities must be conducted to determine detection limits of the species. Unintentional introduction of airborne species can greatly affect the relative read proportions retrieved and confound data interpretation, creation of novel blanks that could be added to field protocols could aid in accounting field and lab-introduced contamination, rather than the occurrence from insect interaction in the environment.
The Project DINA is funded by the German Federal Ministry of Education and Research (Bundesministerium für Bildung und Forschung, BMBF; grant number FKZ 01LC1901).
The authors have declared that no competing interests exist.
Raw sequence data is available on the Sequence Read Archive (SRA) under the accession number PRJNA851235. The R script for the identification database, as well as the database, are in Suppl. material
We are especially thankful to all the volunteers in the field who maintained the Malaise traps. Furthermore, we would like to thank all partners of the DINA consortium for the collaboration and scientific input and especially Wolfgang Wägele for project initiation. Thank you to Sabine Mutz, Kathrin Moses-Luehrsen and Annalena Kurzweil (all University Giessen) for laboratory and administrative support.
Table S1
Data type: table
Explanation note: Generic and species level identification of Amplicon Sequence Variance (ASVs) resulting from ITS2 metabarcoding of Malaise trap plant contents.
Table S2
Data type: table
Explanation note: Species retrieved from ITS2 metabarcoding and vegetation surveys in the area surrounding Malaise traps.
Data type: R script