Assessment of Insect Communities with Metabarcoding

Assessment of Insect Communities with Metabarcoding | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Assessment of Insect Communities with Metabarcoding Hendra F. Sihaloho, Isham Azhar, Millawati Gani, Juliana Senawi, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7367951/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Metabarcoding is widely used for molecular identification of organisms. Due to its efficiency for en masse characterization of samples, the technique is useful for insect biodiversity surveys. Although metabarcoding has been used for nearly two decades, there is still a need for optimized insect sample processing strategies. The goal of this study was to establish best practices for molecular characterization of bulk insect collections. We sampled insect diversity using light traps in lowland dipterocarp forest of Tengku Hassanal Wildlife Reserve in Pahang, Malaysia. Each light trap sample was homogenized and repeatedly subsampled to identify the number of subsamples required to detect total estimated insect OTU diversity in a light trap. Insect OTU diversity from 72 subsamples (12 subsamples from each of six light traps) was characterized by sequencing part of the mitochondrial Cytochrome Oxidase I (COI) gene. We found that four and eight 100 µL subsamples were sufficient to detect at least 90% and 95%, respectively, of insect OTU diversity collected in each light trap sample, regardless of the range of variation in preserved biomass (137.9 ̶ 445.4 g). We also built a global database from the Barcode of Life Database (BOLD) repository for our target sequencing region (313bp). We found that standardized bitscores of tblastx were significantly higher than blastn (p < 0.001); however, the percent identity distributions were not statistically different, with both around 92–93%. In addition to relatively low average match identities, we also found poor taxonomic concordance between blastn and tblastx, especially at lower taxonomic levels, and suggest the advantage of metabarcoding should be leveraged by focusing on phylogenetic diversity instead of taxonomy wherever possible. insect homogenate subsample database sample coverage blast Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Insects are the most speciose animal class, with a million described species and likely at least another 4.5 million (~ 80%) that remain to be discovered (Stork 2018 ). Insects fill essential roles in ecosystems as pollinators, as well as in food webs, and consequently, insect diversity is a valuable measure of forest condition (Macedo-Reis et al. 2019 ; Sánchez-Bayo and Wyckhuys 2019 ). However, studying insect diversity can be challenging due to the vast number of species and individual numbers potentially encountered at a given site. Relatedly, morphology is often used to identify insect samples in surveys; however, this approach has a low throughput, being limited by the requirement for taxonomic expertise and the prevalence of undescribed species (O'Rourke et al. 2021 ). The implementation of molecular approaches over the past few decades, namely, DNA barcoding (Hebert et al. 2003a ; Hebert et al. 2003b ; Ratnasingham and Hebert 2007 ) metabarcoding techniques (Bik et al. 2012 ; Taberlet et al. 2012 ; Yu et al. 2012 ), and the latter method has provided an alternative approach to insect identification that greatly increases throughput. Whereas genetic identification of individual specimens is commonly performed, metabarcoding refers to the sequencing of bulk samples or eDNA containing many individuals and taxa, followed by tabulation of the genetic diversity profiles (Liu et al. 2020 ; Yu et al. 2012 ). Examples of sample types benefiting from metabarcoding include insect diversity surveys (Li et al. 2023 ; Roslin et al. 2022 ; Zenker et al. 2020 ), pest surveillance (Batovska et al. 2021 ; Hardulak et al. 2020 ; Young et al. 2021 ) as well as investigations of insectivore diets by fecal analysis (Alberdi et al. 2020 ; O’Rourke et al. 2022 ; Phillips et al. 2017 ). Bulk samples from insect traps may be prepared for metabarcoding by homogenization prior to DNA extraction. However, traps commonly recover large masses of insects (e.g., from tens to hundreds of grams), but DNA extractions have a practical limit to the amount of input material, so it is not possible to extract from the entire homogenate. One solution is to reduce sample mass by dissecting out legs from individuals to carry forward through extraction (Ji et al. 2013 ). Whereas this approach achieves the goal of reducing input mass, it may be impractical and time-consuming. An alternative approach is sample homogenization followed by subsampling (Zizka et al. 2022 ). Although a sample can theoretically be completely homogenized, in practice it seems unlikely that this will be achieved. Therefore, multiple subsamples are required to recover the diversity collected by the trap. A recent study using Malaise traps in a wetland reserve near Krefeld in Western Germany found that 12–17 subsamples were required to cover 95% of the sample insect diversity (Zizka et al. 2022 ). The homogenization approach employed by Zizka et al. ( 2022 ) represents a valuable contribution to insect metabarcoding techniques, and it is important to characterize the applicability of their recommendations to collections from different ecosystems, such as tropical forests, which have higher insect diversity than temperate regions (Gaston and Hudson 1994 ), and using different trapping methods, for instance, light traps that attract nocturnal flying insects (Shimoda and Honda 2013 ). In addition to sample processing considerations, the interpretation of results in a taxonomic framework based on comparison of identified sequences to a reference database is a common component of analysis and inference (Magoga et al. 2022 ). A few technical issues potentially limit accurate taxonomic assignment. First, despite great effort (i.e., Barcode of Life Database (BOLD); Ratnasingham and Hebert 2007 ), representative DNA sequences for most of the world’s insect taxa are not currently available, and will not be so for the foreseeable future. Second, due to sequencing and bioinformatic technical constraints, metabarcoding approaches have relied on relatively short mitogenomic region as the genetic barcode locus. Thus, database preparation and taxonomic interpretation should carefully consider taxonomic reliability. We sampled insects using light traps on the Tengku Hassanal Wildlife Reserve (THWR) in Pahang, Malaysia. The THWR, recently renamed from Krau Wildlife Reserve, is the third largest protected area in Peninsular Malaysia (Zakaria et al. 2014 ). Like other areas in Southeast Asia, the area around THWR has been extensively converted to non-timber plantations (rubber Hevea brasiliensis , oil palm Elaeis guineensis ), resulting in a mosaic landscape of various land uses and forest fragments. The goal of this work was to contribute to the body of best practice for characterizing insect diversity by bulk sampling and metabarcoding, specifically by sampling in a tropical rainforest with light traps. We sought to estimate: (1) the number of subsamples from homogenates required to recover at least 90% sample diversity; (2) the number of sampling nights required to detect 90% of insect OTU diversity at a locality over time; (3) evaluate the best approach and limitations of taxonomic inference. Materials and Methods Study Sites Sampling was conducted from May 26th to July 3rd, 2019, at Tengku Hassanal Wildlife Reserve in Pahang, Malaysia (THWR; Table 1 ; 50m asl). Those three months were previously reported to have peak insect biomass and rainfall in 2009 and 2010 (Nurul-Ain et al. 2017 ). The reserve covers an area of 60,349 hectares of continuous dipterocarp forest under the management of the Department of Wildlife and National Parks (PERHILITAN). The average annual rainfall of the area is about 2,000 mm, and the daily temperature ranges from 23 o C to 33 o C (Zakaria et al. 2014 ). Insect trapping was conducted for six nights at six different locations (Supp.Table 1 ) at Kuala Lompat Research Station, which is located in the southeastern part of the reserve. Table 1 Light trap samples: Sample ID, Subsample ID, Sampling dates, GPS coordinates, and the wet weight. Sample ID Subsample ID Sampling Date Latitude ( o N) Longitude ( o E) Wet weight (g) TK05 TK05A - TK05L 03 Jul 2019 3.7134 102.2739 165.1 TK06 TK06A - TK06L 20 Jun 2019 3.7177 102.2763 137.9 TK07 TK07A - TK07L 22 Jun 2019 3.7182 102.2737 445.4 TK08 TK08A - TK08L 06 Jun 2019 3.7150 102.2786 197.9 TK09 TK09A - TK09L 27 Jun 2019 3.7157 102.2759 185.1 TK10 TK10A - TK10L 26 May 2019 3.7151 102.2821 157.6 mean ± standard deviation 214.83 ± 114.89 Trapping Methods Insects were collected using light traps (Leptraps LLC) deployed across a 1 km 2 area of THWR, and the median distance between traps was 531 m. Traps were equipped with an 18-inch 12V 32-watt quantum black light bulb with rigid plexiglass vanes, a photoelectric switch, and a funnel placed at the opening of a plastic bucket containing Propylene Glycol (PG) (Moreau et al. 2013 ; Patrick et al. 2016 ) for preserving collected insects. Light traps were powered by 12V 20AH sealed lead acid batteries. They were deployed daily around 6.30pm (dusk) and collected around 6.30am (dawn) the following morning. Insect samples from each trap were poured into a collection jar labelled with trap number, date, and time of collection. The insect jars were kept at ambient temperature before being processed in the laboratory. Sample Processing Contents of each insect jar were processed separately, starting by pouring insects onto a 0.8 mm wire mesh stainless steel sieve (Hallmann et al. 2017 ). Sieving continued until drainage droplets were separated by 10 seconds or more. The sieved sample was then weighed using a scale with an accuracy of at least 0.1 g to obtain bulk biomass, which we refer to hereafter as wet weight. The sample was transferred to a cleaned blender (Model 7010S, Waring Commercial) for homogenization. The homogenization was carried out by blending the insect samples with 50 mL digestion buffer (pH 8.0) for 90 seconds. Fresh digestion buffer was prepared before processing insect samples using 10 mM Tris-HCL, 10 mM NaCl, 5 mM CaCl 2 , 2.5 mM EDTA, and 2‰ SDS (Campos and Gilbert 2012 ). Following the homogenization, 25 mL of the homogenate was transferred into 50 mL falcon tubes and stored at -80 o C. Each homogenate was subsampled 12 times by thawing the samples in an ice bucket, mixing thawed homogenate using a vortex, then using a blunted 1 mL pipet tip to pipet 100 µL of homogenate into 2 mL sterile vials containing 1 mL of fresh digestion buffer, this was the same digestion buffer previously described with additional 50mM DTT and 100 µg proteinase K. Homogenate was re-vortexed between each 100 µL subsample. The subsample vials were incubated in 50 o C water bath for at least 18 hours before they were stored at -80 o C. DNA was extracted from each subsample using DNeasy Blood & Tissue Kit (Qiagen Inc.) following the manufacturer’s protocol. Each DNA extract was assessed for quality and quantity using Nanodrop (Thermo Fisher Scientific Inc.). Cytochrome Oxidase I Amplification and Sequencing Metabarcoding was employed to assign the taxonomy of insect Operational Taxonomic Units (OTU). A region of Cytochrome Oxidase I (COI) was amplified using mlCOIintF (Leray et al. 2013 ) and HCO2198R (Folmer et al. 1994 ) primers that targeted a 313 bp amplicon length. Both primers were designed to include the Illumina 5' overhang, so that the Illumina barcodes could be directly added to the purified PCR products in the second round of amplification. The forward and reverse primers used in the first round PCR were TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG GGWACWGGWTGAACWGTWTAYCCYCC and GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG TAAACTTCAGGGTGACCAAAAAATCA (the sequences of mlCOIintF and HCO2198R primers, respectively, are boldface). The first round of PCR reactions was carried out in 25 µL total volumes containing 2.5 µL DNA template (~ 3ng), 1X Q5 Hifi Hotstart ReadyMix (New England Biolabs Inc.), and 0.2 µM of each primer. The following temperature profile was used to amplify the COI region: A 5min initial denaturation at 98 o C, followed by 25 cycles of 30s denaturation at 98 o C, 40s annealing at 54 o C, 1min extension at 72 o C, and a final extension at 72 o C for 10min. PCR products were purified using 0.8X AMPure XP beads (Beckman Coulter, Inc., IN, USA). The second round of amplifications was performed to incorporate the Illumina i5 and i7 adapters and 8bp barcodes. These reactions were performed in a volume of 10 µL, containing 5 µL of KAPA HiFi HotStart Ready Mix (Roche Diagnostics), 1 µL each of i5 and i7 barcoded adapters at 5 pM, and 3 µL of purified PCR products. The thermal profile for the second PCR was 3min initial denaturation at 95 o C, followed by 8 cycles of 30s denaturation at 95 o C, 30s annealing at 55 o C, 30s extension at 72 o C, and a final extension at 72 o C for 1min. Libraries were purified using 0.7X AMPure XP beads (Beckman Coulter, Inc., IN, USA). Then, libraries were analyzed using a Qubit fluorometer (Qubit dsDNA HS Assay Kit; Life Technologies, CA, USA) and TapeStation (High Sensitivity D1000 ScreenTape assay; Agilent Technologies, Inc., CA, USA) to determine concentrations and fragment sizes, respectively, and subsequently pooled in equimolar amounts. The pooled libraries were sequenced on a MiSeq platform using 2 x 300 bp sequencing and MiSeq Reagent Kit v3 (Illumina, Inc.) at the Monash University Malaysia Genomics Platform. Building the insect COI reference database We developed a database from the BOLD repository (Fig. 1 ). Insect sequences were downloaded from BOLD (Ratnasingham and Hebert 2007 ) in December 2022 using modified R scripts by O'Rourke et al. ( 2021 ) in RStudio (Posit team 2023 ; R Core Team 2023 ). Our database pipeline was modified from previously described approaches (O'Rourke et al. 2020 ; Robeson et al. 2021 ) and included removal of any leading and trailing Ns (uncalled bases), followed by removal of homopolymers equal to or greater than 12 base pairs and ambiguous nucleotide runs equal to or greater than 6 base pairs. Obtaining good starting alignment of millions of sequences is computationally challenging, so we first reduced the database to Malaysian entries, which were subset to 655–1500 bp length entries and aligned using MAFFT v7.490 (Katoh and Standley 2013 ). The mlCOIintF and HCO2198R primers (Folmer et al. 1994 ; Leray et al. 2013 ) were then added to this alignment to verify coordinates and the length of the targeted sequencing region. Next, the remaining database entries were added to the alignment using MAFFT v7.490 (Katoh and Standley 2013 ) with –auto –addfull flags. The aligned region of the database was extracted using extract_alignment_region.py script (Robeson et al. 2021 ), and sequences were then trimmed to the COI region flanked by primers described above (300–319 bp) using SeqKit v2.3.0 (Shen et al. 2016 ). The Nuclear mitochondrial DNA segments (NUMTs) were then removed using the Biostrings v2.72.1 package in R (Pagès et al. 2024 ). Next, the unique sequences were identified with Usearch v11 --fastx_uniques command (Edgar 2010 ). Any unique sequences represented by multiple sequences with conflicting taxonomy were removed. Processing of COI metabarcoding data Metabarcoding sequence data were processed following Phillips et al. ( 2017 ; Fig. 1 ). Briefly, primers were removed from raw reads using cutadapt v2.6 (Martin 2011 ). Paired-end reads were merged using PEAR v0.9.11 with default settings (Zhang et al. 2014 ). Usearch v11 (Edgar 2010 ) was employed for quality filtering, identifying and abundance counting of unique sequences, and clustering based on 97% similarity to inform OTU table construction. A multiple sequence alignment of the representative OTU sequences was created by first using MAFFT v7.490 (Katoh and Standley 2013 ) with the –auto flag applied to a subsample of 200 sequences. Next, the remaining representative OTU sequences were added and aligned with –auto –addfull flags. Any OTUs with a length less than 300 were then removed using SeqKit v2.3.0 (Shen et al. 2016 ). The OTUs containing NUMTs were removed using Biostrings v2.72.1 R package (Pagès et al. 2024 ). Data management was carried out using the phyloseq v1.48 R package (McMurdie and Holmes 2013 ) in RStudio (Posit team 2023 ; R Core Team 2023 ). Taxonomy for each OTU was independently assigned using blastn and tblastx (v2.14.1+; Camacho et al. 2009 ) with an e-value threshold of 1e-4. For each OTU, we retained the top hit based on bit score value. Both blastn and tblastx were evaluated to understand if nucleotide- versus protein-level comparison offered advantages. The top bit score, which is a widely used measure of alignment significance, was recorded for each OTU for each database. Bit score was selected as the comparator metric because it is comparable across databases of different sizes and more reliable than percent identity for inferring homology (Pearson 2013 ). Since the values of bitscore from blastn and tblastx had different possible ranges, each was transformed using the proportion of maximum (POM) before comparison using the Wilcoxon signed rank test (Quinn and Keough 2002 ; Zar 2010 ). Biological diversity and sample coverage Hill numbers (0, 1, 2) were calculated to summarize insect OTU diversity detected from subsamples and for entire light traps (Alberdi and Gilbert 2019 ; Chao et al. 2014 ). Unbiased comparison of communities requires parity of sample coverage, defined as “ the proportion of the total number of individuals in a community that belong to the species represented in the sample (Chao and Jost 2012 ). Sample coverage estimates how well true diversity is obtained by a sample (Roswell et al. 2021 ). One way to approach parity is to set a coverage threshold and then compare communities at this threshold. In this study, we wanted to determine the number of subsamples that would recover at least 90% and 95% of insect OTU diversity for each sample. Sample coverage was estimated at two levels: (1) the number of subsamples per light trap homogenate required to recover insect OTU diversity for each light trap, and (2) the number of sampling nights needed to characterize insect OTU diversity at a locality. Sample coverage was calculated using the iNEXT v2.0.20 R package (Chao et al. 2014 ; Hsieh et al. 2016 ). The OTU table was normalized to the lowest sample-wide sequencing depth using stratified random sampling (Beule and Karlovsky 2020 ). Then, sample coverage was calculated to determine the number of subsamples required to represent at least 90% and 95% coverage for each sample. Each light trap sample was subsampled 12 times, and extrapolation was limited to 24 subsamples to avoid estimation bias (Chao et al. 2014 ; Chao and Jost 2012 ). Also, to evaluate the efficacy of homogenization, one-way Analysis of Variance (ANOVA) was used to test if subsample combination affected diversity recovered (subsample combination at a given subsampling depth was a nested variable within light trap). At the locality-level, sample coverage was calculated to determine the number of light traps required to represent at least 90% and 95% of local diversity. Analyses in R also employed the following packages: bold v1.3.0 (Dubois and Chamberlain 2023 ), SRS v0.23 (Beule and Karlovsky 2020 ; Heidrich et al. 2021 ), taxize v0.9.100 (Chamberlain and Szöcs 2013 ; Chamberlain et al. 2020 ), tidyverse v2.0.0 (Wickham et al. 2019 ), VennDiagram v1.7.3 (Chen 2022 ), and reshape2 (Wickham 2007 ). Results Insect OTU diversity A total of 3,329,140 paired-end reads were obtained study-wide with 23,119 ± 2,611 (mean ± sd) per subsample. The number of reads retained through bioinformatic processing across subsamples ranged from 15,788 to 28,222, with an average of 20,615 ± 2,524. The range in wet weight from the six light traps was 137.9 to 445.4 g (average 214.83 ± 114.89 g; Table 1 ), from which 3,058 OTUs at the 97% similarity level were discovered. The average number of insect OTUs per trap subsample was 508 ± 136. An analysis of variance (ANOVA) was conducted to assess the efficacy of homogenization, and it indicated no significant effect of subsample combination on insect OTU richness (p > 0.05; Table S1 ). Reference database A total of 4,241,641 database entries were downloaded from BOLD using “insecta” as search query criteria and filtered for marker codes “COI-5P”. Table 2 summarizes the number of starting and ending database entries retained at each step in the database development workflow. It showed that the final database only preserved 17.14% of the original database entries (726,899 sequences). Table 2 Number of candidate database sequence entries (class insecta) retained at each step of the database development workflow. Development Steps Number of entries Retain (%) Raw entries 4,241,641 100.00 Removing homopolymer 4,229,603 99.72 Trimming to the sequenced region 2,484,568 58.58 Remove NUMTs 2,453,660 57.85 Retain unique sequences 745,041 17.56 Remove taxonomic discrepancies 726,899 17.14 For the total 3,058 insect OTUs, blastn analysis returned significant hits for 3,056 OTUs, while tblastx analysis provided significant matches for all OTUs for each database. The median nucleotide (blastn) and translated protein (tblastx) match identities were 92.01% and 93.20%. Comparison of standardized bit scores between blastn and tblastx revealed that tblastx was significantly better (p 0.05). The number of insect OTUs with significant matches to either blastn or tblastx searches varied across taxonomic levels (Fig. 2 ). In general, the concordance of taxonomic assignment between blastn and tblastx was smaller at lower taxonomic levels. Out of 3,058 insect OTUs, only 518 (16.94%) had concordant species-level identifications between blastn and tblastx. This number increased to 878 OTUs (28.71%) at the generic level, 2,045 (66.87%) at the familial level, and 2,858 (93.46%) at the ordinal level. A similar trend was also observed when a minimum 85% identity parameter was applied to the best blastn or tblastx hits (Fig. 2 , bottom row). Although it was not significant, the number of OTUs with concordant taxonomic assignments decreased when an 85% PID was applied. Light Trap Diversity and Completeness The diversity of insect OTUs contained in each of the 12 subsamples for each of six light traps was summarized by a Hill 0 (richness) of 902.5 ± 243.05, Hill 1 (exponential Shannon-Wiener) 689.59 ± 199.44, and Hill 2 (Inverse Simpson) 625.27 ± 183.01. As expected, increasing the number of subsamples for a given light trap enhanced the estimated sample completeness (Fig. 4 ). All samples, except TK06, reached 98% sample coverage when using 12 subsamples (98.13 ± 1.00); light trap TK06 only reached 96.2% sample coverage when using 12 subsamples, but TK06 was not an outlier with respect to wet weight (Table 1 ). With the intent to identify the minimum number of required subsamples to detect at least 90% insect OTUs diversity, four subsamples per light trap achieved this threshold, averaging 93.83% ± 1.62. For comparison, a minimum of 8 subsamples was required to detect at least 95% insect OTU diversity in a light trap. The six insect light traps studied here came from six different nights spanning 39 days (May 26th to July 3rd, 2019) within a forest habitat < 1 km 2 . Using four subsamples per light trap, the number of traps needed to discover insect OTU diversity in the given spatiotemporal setting was investigated. The diversity of insect OTU for six nights was 2,701 (Hill 0 ), 2,267 (Hill 1 ), and 1,845 (Hill 2 ). Results indicated that six sampling nights only covered 59.8% of insect OTU diversity (Fig. 5). Extrapolation, which has estimation uncertainty due to needing to extrapolate greater than double the number of actual sample nights, indicated that 21 and 28 trap nights would be required to reach 90% and 95% OTU diversity of nocturnal flying insects, respectively. Discussion Studies of insect diversity generally follow two approaches: non-destructive and destructive sampling. The non-destructive methods, such as morphological identification, ethanol-based DNA extraction of parts of individual specimens, and imaging prior to DNA extraction, offer the key advantage of preserving physical specimens for future examination, taxonomic validation, or voucher deposition. However, these methods are often more labor-intensive and may not scale efficiently for large biodiversity assessments. In contrast, destructive metabarcoding approaches, such as the homogenization of bulk samples (Yu et al. 2012 ), prioritize high-throughput processing and standardization. In our current study, we adopted the homogenization approach with the goal of optimizing the number of subsamples required to effectively capture the majority of molecular operational taxonomic units (OTUs) within a bulk insect sample. While this method sacrifices the preservation of individual specimens, it provides a practical and efficient solution for broad-scale biodiversity assessment. Insect diversity and reference database A total of 3,058 97% insect OTUs were identified in this study from 6 light traps collected on separate nights across a 39-day span within a 1 km 2 forest area in peninsular Malaysia. Total diversity reported herein is higher than that reported in a study of insect pests in diverse biomes of Brazil (Zenker et al. 2020 ) as well as one from a German wetland (Zizka et al. 2022 ). Many differences among studies, including sampling duration, geographic spread, trapping method, molecular and analytical choices, preclude direct comparisons. However, given both the large volume of insects attracted by the light traps used herein, and the high biodiversity of the study system, the provided recommendations on sample treatment are likely to be applicable across many different study systems. Also, there is no reason to expect that the challenges encountered with taxonomic assignment are unique to this study. There is an estimated high percentage of insect biodiversity for which genetic representatives are lacking. Given the disparity between what is known and what will be encountered in a metabarcoding study at a given site, it seems intuitive that comparing observed genetic lineages to all available information is the best option for taxonomic assignment. Still, database preparation herein revealed that for the COI gene region considered, there were instances of multiple BOLD entries having identical sequences (~ 1.7 million) as well as entries excluding the region targeted by the assay (~ 1.7 million). In total, about 3.4 million sequences, equating to about 40% of sequences considered for database inclusion, were identified on these criteria. Representing such sequence duplicates as a single representative sequence with taxonomy trimmed back to the lowest congruent level serves to both reduce database size and reflect taxonomic uncertainty. Insect OTUs were compared to the reference database using both nucleotide (blastn) and protein (tblastx) based approaches, with the rationale being to assess if queries leveraging more (nucleotide) or less (protein) variable sequences offered any advantages. From direct comparison, relatively high taxonomic concordance was observed at the ordinal level, however, low concordances elsewhere (familial (68%), genera (41%), and species (29%) levels). Also, average percentage identities for nucleotide- and protein-level comparisons did not significantly differ and were, at around 92–93%, below that commonly expected for intraspecific comparisons. The general taxonomic uncertainty is thought to be caused not only by a lack of sequence representatives for much Insecta diversity, but also due to a lack of species-level resolution influenced by homoplasy in a small genomic region. Sample Completeness A main goal herein was to understand effort requirements to capture 90 or 95% insect OTU diversity in homogenate samples. Therefore, subsequent investigations into requirements for light trap subsampling and field efforts were conducted using 97% OTUs as the level of evolutionary delineation. Rarefaction curves (Fig. 3 ) suggested that additional effort beyond four subsamples would, on average, only cover an additional 6% of diversity disproportionately comprised of rare OTUs. For instance, for obtaining 95% sample coverage, a total of eight subsamples was required. Whereas in this study, each subsample was amplified and sequenced separately to allow investigation of the subsampling procedure, future work following these guidelines can pool subsamples’ DNA after extraction and prior to sequencing for cost savings. Thus, achieving 95% coverage would not require much expense beyond that required for 90% sample coverage. That said, pooling subsamples could introduce PCR biases, so treating subsamples independently throughout the process is preferred if cost is not prohibitive. Other recent work by Zizka et al. ( 2022 ) was structured similarly to the current study but used Malaise traps and conducted field work in a wetland nature reserve near Krefeld in western Germany. That study reported that 12–17 subsamples were required to capture 95% diversity, a level of diversity detected by about eight subsamples in the current study. This is a surprising contrast, especially given that tropical rainforests support more insect diversity than temperate regions (Lewinsohn and Roslin 2008 ). Multiple aspects of their study design and analytical approach differed from those reported herein, which could explain this difference. One possible key difference between the two studies is the amount of sequencing effort per subsample; the current study had an average of 23,119 reads per subsample, whereas Zizka et al. ( 2022 ) reported an average of 154,659 reads per subsample. It is simultaneously possible that the 7-fold increased sequencing effort by Zizka et al. ( 2022 ) allowed greater detection of rare OTUs, but also lead to the creation of erroneous OTUs through recurrent sequencing error across great sequencing depth. The latter is a real possibility given Illumina’s average error rate of 1 out of 1000 base calls in conjunction with deep sequencing of PCR amplicons. It is not possible to delineate these effects, but re-analysis of the Zizka et al. ( 2022 ) data set using the bioinformatic approach reported herein with sequence data first subsampled to the current studies’ average sequencing effort also resulted in four subsamples detecting at least 90% (93.82 ± 0.99) total diversity across all samples (Figure S1 , Table S2). Thus, a reasonable joint interpretation appears to be that four subsamples is a generalizable recommendation to achieve a minimum 90% sample coverage so long as sequencing effort is regulated, and additional subsampling will provide more but modest additional information. For purposes of sample coverage estimation, extrapolation greater than two times the number of actual samples has been cautioned because of increased bias (Chao et al. 2014 ). Nevertheless, to provide some insight into how much trapping effort would be required to summarize insect OTU diversity at the study site adequately, it was estimated that 21 sampling nights, with one trap set per night, would be required to reach 90% completeness. This is with respect to the 39 day and 1 km 2 spatiotemporal interval from which the 6 samples were obtained. Our interpretation is that insect biodiversity turnover is quite high in the Malaysian rainforest study site, and with respect to the spatial scale of the current study, trapping would need to occur about every other night. This requirement would obviously change depending on spatiotemporal scale. Interesting future work possibilities include longitudinal monitoring of numerous set trapping stations across the landscape. Conclusion Optimizing resources (funds, time, personnel) is crucial to the design of ecological fieldwork, especially in rapid biodiversity assessments, such as monitoring or large-scale biodiversity projects. Results demonstrate the feasibility of the reported light trap homogenization strategy and suggest that four to eight subsamples of homogenate should detect the large majority of insect OTU diversity in bulk collections. Contrasts with other similar work suggest that limiting sequencing depth is an important consideration to avoid the creation of false OTUs due to excessive sequencing effort. Finally, given the high insect diversity in tropical forests, trapping for approximately 20 nights within a roughly 1 km² area over a one-month time frame is expected to facilitate the detection of the large majority of nocturnal flying insect diversity. Studies moving forward should verify subsampling species accumulation wherever possible, but we expect these suggestions to be generally applicable. Declarations Acknowledgement The authors thank the Economic Planning Unit of the Prime Minister’s Department (EPU 40/200/19/3510), Department of Wildlife and National Parks (PERHILITAN) Peninsular Malaysia (JPHL & TN (IP): 100-34/1.24 Jld. 14 (31), and Access & Benefit Sharing - Pahang Biodiversity Council (MBP.600(S)-1/1/1 Jld. 4 (10) for permission to conduct research in Malaysia. We thank the Natural Science Research Laboratory, Museum of Texas Tech University, for assistance with field supplies and organization, and Sze Mei Lee for her technical assistance in preparing some of the DNA libraries. This material is based upon work supported by the National Science Foundation under Grant No. DEB-1754810. LSY and QA were supported by a strategic grant from the School of Science, Monash University Malaysia (Grant No I-M010-STG-000188). Conflict of Interest The authors declare no conflict of interest. Authors' Contributions This paper was completed through collaboration among all authors. TK and CDP designed and acquired funding for the project. IA, TK, JS and CDP carried out the fieldwork. MG, LSY, and QA conducted laboratory procedures. HFS, CDP, IA, TK and RD analyzed the data. HFS & CDP wrote the first draft of the manuscript. All authors reviewed the manuscript and provided critical advice. Data Availability Statement The raw sequence data generated in this study are available at GenBank Sequence Read Archive, BioProject accession number PRJNA1260500. All scripts of the current study can be accessed at https://github.com/mhenso/insect_metabarcoding. References Alberdi A, Gilbert MTP (2019) A guide to the application of Hill numbers to DNA-based diversity analyses. Mol Ecol Resour 19(4):804–817. https://doi.org/10.1111/1755-0998.13014 Alberdi A, Razgour O, Aizpurua O, Novella-Fernandez R, Aihartza J, Budinski I, Garin I, Ibáñez C, Izagirre E, Rebelo H, Russo D, Vlaschenko A, Zhelyazkova V, Zrnčić V, Gilbert MTP (2020) DNA metabarcoding and spatial modelling link diet diversification with distribution homogeneity in European bats. Nat Commun 11(1):1154. https://doi.org/10.1038/s41467-020-14961-2 Batovska J, Piper AM, Valenzuela I, Cunningham JP, Blacket MJ (2021) Developing a non-destructive metabarcoding protocol for detection of pest insects in bulk trap catches. Sci Rep 11(1):7946. https://doi.org/10.1038/s41598-021-85855-6 Beule L, Karlovsky P (2020) Improved normalization of species count data in ecology by scaling with ranked subsampling (SRS): application to microbial communities. PeerJ 8:e9593. https://doi.org/10.7717/peerj.9593 Bik HM, Porazinska DL, Creer S, Caporaso JG, Knight R, Thomas WK (2012) Sequencing our way towards understanding global eukaryotic biodiversity. Trends Ecol Evol 27(4):233–243. https://doi.org/10.1016/j.tree.2011.11.010 Camacho C, Coulouris G, Avagyan V, Ma N, Papadopoulos J, Bealer K, Madden TL (2009) BLAST+: architecture and applications. BMC Bioinformatics 10(1):421. https://doi.org/10.1186/1471-2105-10-421 Campos PF, Gilbert TMP (2012) DNA extraction from keratin and chitin. In: Shapiro B, Hofreiter M (eds) Ancient DNA: methods and protocols. Humana Press, Totowa, NJ, pp 43–49 Chamberlain SA, Szöcs E (2013) taxize: taxonomic search and retrieval in R. F1000Research 2(191). https://doi.org/10.12688/f1000research.2-191.v2 Chamberlain SA, Szoecs E, Foster Z, Arendsee Z, Boettiger C, Ram K, Bartomeus I, Baumgartner J, O'Donnell J, Oksanen J, Tzovaras BG, Marchand P, Tran V, Salmon M, Li G, Grenié M (2020) taxize: Taxonomic information from around the web Chao A, Gotelli NJ, Hsieh TC, Sander EL, Ma KH, Colwell RK, Ellison AM (2014) Rarefaction and extrapolation with Hill numbers: a framework for sampling and estimation in species diversity studies. Ecol Monogr 84(1):45–67. https://doi.org/10.1890/13-0133.1 Chao A, Jost L (2012) Coverage-based rarefaction and extrapolation: standardizing samples by completeness rather than size. Ecology 93(12):2533–2547. https://doi.org/10.1890/11-1952.1 Chen H (2022) VennDiagram: generate high-resolution Venn and Euler plots Dubois S, Chamberlain S (2023) bold: Interface to Bold Systems API Edgar RC (2010) Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26(19):2460–2461. https://doi.org/10.1093/bioinformatics/btq461 Folmer O, Black M, Hoeh W, Lutz R, Vrijenhoek R (1994) DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Mol Mar Biol Biotechnol 3(5):294–299 Gaston KJ, Hudson E (1994) Regional patterns of diversity and estimates of global insect species richness. Biodivers Conserv 3(6):493–500. https://doi.org/10.1007/BF00115155 Hallmann CA, Sorg M, Jongejans E, Siepel H, Hofland N, Schwan H, Stenmans W, Müller A, Sumser H, Hörren T, Goulson D, de Kroon H (2017) More than 75 percent decline over 27 years in total flying insect biomass in protected areas. PLoS One 12(10):e0185809. https://doi.org/10.1371/journal.pone.0185809 Hardulak LA, Morinière J, Hausmann A, Hendrich L, Schmidt S, Doczkal D, Müller J, Hebert PDN, Haszprunar G (2020) DNA metabarcoding for biodiversity monitoring in a national park: screening for invasive and pest species. Mol Ecol Resour 20(6):1542–1557. https://doi.org/10.1111/1755-0998.13212 Hebert PDN, Cywinska A, Ball SL, deWaard JR (2003a) Biological identifications through DNA barcodes. Proc R Soc Lond B Biol Sci 270(1512):313–321. 10.1098/rspb.2002.2218 Hebert PDN, Ratnasingham S, de Waard JR (2003b) Barcoding animal life: cytochrome c oxidase subunit 1 divergences among closely related species. Proc R Soc Lond B Biol Sci 270(Suppl 1):S96–S99. https://doi.org/10.1098/rsbl.2003.0025 Heidrich V, Karlovsky P, Beule L (2021) ‘SRS’ R Package and ‘q2-srs’ QIIME 2 Plugin: Normalization of Microbiome Data Using Scaling with Ranked Subsampling (SRS) Applied Sciences, vol. 11 Hsieh TC, Ma KH, Chao A (2016) iNEXT: an R package for rarefaction and extrapolation of species diversity (Hill numbers). Methods Ecol Evol 7(12):1451–1456. https://doi.org/10.1111/2041-210X.12613 Ji Y, Ashton L, Pedley SM, Edwards DP, Tang Y, Nakamura A, Kitching R, Dolman PM, Woodcock P, Edwards FA, Larsen TH, Hsu WW, Benedick S, Hamer KC, Wilcove DS, Bruce C, Wang X, Levi T, Lott M, Emerson BC, Yu DW (2013) Reliable, verifiable and efficient monitoring of biodiversity via metabarcoding. Ecol Lett 16(10):1245–1257. https://doi.org/10.1111/ele.12162 Katoh K, Standley DM (2013) MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol Biol Evol 30(4):772–780. http://dx.doi.org/10.1093/molbev/mst010 Leray M, Yang JY, Meyer CP, Mills SC, Agudelo N, Ranwez V, Boehm JT, Machida RJ (2013) A new versatile primer set targeting a short fragment of the mitochondrial COI region for metabarcoding metazoan diversity: application for characterizing coral reef fish gut contents. Frontiers in Zoology 10(1):34. https://doi.org/10.1186/1742-9994-10-34 Lewinsohn TM, Roslin T (2008) Four ways towards tropical herbivore megadiversity. Ecol Lett 11(4):398–416. https://doi.org/10.1111/j.1461-0248.2008.01155.x Li M, Lei T, Wang G, Zhang D, Liu H, Zhang Z (2023) Monitoring insect biodiversity and comparison of sampling strategies using metabarcoding: a case study in the Yanshan Mountains, China. Ecol Evol 13(4):e10031. https://doi.org/10.1002/ece3.10031 Liu M, Clarke LJ, Baker SC, Jordan GJ, Burridge CP (2020) A practical guide to DNA metabarcoding for entomological ecologists. Ecol Entomol 45(3):373–385. https://doi.org/10.1111/een.12831 Macedo-Reis LE, Quesada M, de Siqueira Neves F (2019) Forest cover drives insect guild diversity at different landscape scales in tropical dry forests. For Ecol Manage 443:36–42. https://doi.org/10.1016/j.foreco.2019.04.007 Magoga G, Forni G, Brunetti M, Meral A, Spada A, De Biase A, Montagna M (2022) Curation of a reference database of COI sequences for insect identification through DNA metabarcoding: COins. Database 2022:baac055. https://doi.org/10.1093/database/baac055 Martin M (2011) Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnetjournal; Vol 17, No 1: Next Generation Sequencing Data Analysis. https://doi.org/10.14806/ej.17.1.200 McMurdie PJ, Holmes S (2013) phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS One 8(4):e61217. https://doi.org/10.1371/journal.pone.0061217 Moreau CS, Wray BD, Czekanski-Moir JE, Rubin BER (2013) DNA preservation: a test of commonly used preservatives for insects. Invertebrate Systematics 27(1):81–86. https://doi.org/10.1071/IS12067 Nurul-Ain E, Rosli H, Kingston T (2017) Resource availability and roosting ecology shape reproductive phenology of rain forest insectivorous bats. Biotropica 49(3):382–394. https://doi.org/10.1111/btp.12430 O'Rourke DR, Bokulich NA, Jusino MA, MacManes MD, Foster JT (2020) A total crapshoot? evaluating bioinformatic decisions in animal diet metabarcoding analyses. Ecol Evol 10(18):9721–9739. https://doi.org/10.1002/ece3.6594 O'Rourke DR, Mangan MT, Mangan KE, Bokulich NA, MacManes MD, Foster JT (2021) Lord of the diptera (and moths and a spider): molecular diet analyses and foraging ecology of Indiana bats in Illinois. Frontiers in Ecology and Evolution 9. https://doi.org/10.3389/fevo.2021.623655 O’Rourke D, Rouillard NP, Parise KL, Foster JT (2022) Spatial and temporal variation in New Hampshire bat diets. Sci Rep 12(1):14334. https://doi.org/10.1038/s41598-022-17631-z Pagès H, Aboyoun P, Gentleman R, DebRoy S (2024) Biostrings: Efficient manipulation of biological strings Patrick HJH, Chomič A, Armstrong KF (2016) Cooled propylene glycol as a pragmatic choice for preservation of DNA from remote field-collected diptera for next-generation sequence analysis. J Econ Entomol 109(3):1469–1473. https://doi.org/10.1093/jee/tow047 Pearson WR (2013) An introduction to sequence similarity (“homology”) searching. Current Protocols in Bioinformatics 42(1):3.1.1–3.1.8. https://doi.org/10.1002/0471250953.bi0301s42 Phillips CD, Hanson J, Wilkinson JE, Koenig L, Rees E, Webala P, Kingston T (2017) Microbiome structural and functional interactions across host dietary niche space. Integr Comp Biol 57(4):743–755. https://doi.org/10.1093/icb/icx011 Posit team (2023) RStudio: integrated development environment for R. Posit Software, PBC., Boston, MA Quinn GP, Keough MJ (2002) Experimental design and data analysis for biologists. Cambridge University Press, Cambridge R Core Team (2023) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria Ratnasingham S, Hebert PDN (2007) bold: the barcode of life data system (http://www.barcodinglife.org). Mol Ecol Notes 7(3):355–364. https://doi.org/10.1111/j.1471-8286.2007.01678.x Robeson MS, II, O’Rourke DR, Kaehler BD, Ziemski M, Dillon MR, Foster JT, Bokulich NA (2021) RESCRIPt: Reproducible sequence taxonomy reference database management. PLoS Comput Biol 17(11):e1009581. https://doi.org/10.1371/journal.pcbi.1009581 Roslin T, Somervuo P, Pentinsaari M, Hebert PDN, Agda J, Ahlroth P, Anttonen P, Aspi J, Blagoev G, Blanco S, Chan D, Clayhills T, deWaard J, deWaard S, Elliot T, Elo R, Haapala S, Helve E, Ilmonen J, Hirvonen P, Ho C, Itämies J, Ivanov V, Jakovlev J, Juslén A, Jussila R, Kahanpää J, Kaila L, Jari P, Kakko A, Kakko I, Karhu A, Karjalainen S, Kjaerandsen J, Koskinen J, Laasonen EM, Laasonen L, Laine E, Lampila P, Levesque-Beaudin V, Lu L, Lähteenaro M, Majuri P, Malmberg S, Manjunath R, Martikainen P, Mattila J, McKeown J, Metsälä P, Miklasevskaja M, Miller M, Miskie R, Muinonen A, Veli M, Naik S, Nikolova N, Nupponen K, Ovaskainen O, Österblad I, Paasivirta L, Pajunen T, Parkko P, Paukkunen J, Penttinen R, Perez K, Pohjoismäki J, Prosser S, Raekunnas M, Rahulan M, Rannisto M, Ratnasingham S, Raukko P, Rinne A, Rintala T, Miranda Romo S, Salmela J, Salokannel J, Savolainen R, Schulman L, Sihvonen P, Soliman D, Sones J, Steinke C, Ståhls G, Tabell J, Tiusanen M, Várkonyi G, Vesterinen EJ, Viitanen E, Vikberg V, Viitasaari M, Vilen J, Warne C, Wei C, Winqvist K, Zakharov E, Mutanen M (2022) A molecular-based identification resource for the arthropods of Finland. Mol Ecol Resour 22(2):803–822. https://doi.org/10.1111/1755-0998.13510 Roswell M, Dushoff J, Winfree R (2021) A conceptual guide to measuring species diversity. Oikos 130(3):321–338. https://doi.org/10.1111/oik.07202 Sánchez-Bayo F, Wyckhuys KAG (2019) Worldwide decline of the entomofauna: a review of its drivers. Biol Conserv 232:8–27. https://doi.org/10.1016/j.biocon.2019.01.020 Shen W, Le S, Li Y, Hu F (2016) SeqKit: A cross-platform and ultrafast toolkit for fasta/q file manipulation. PLoS One 11(10):e0163962. https://doi.org/10.1371/journal.pone.0163962 Shimoda M, Honda K-i (2013) Insect reactions to light and its applications to pest management. Applied Entomology and Zoology 48(4):413–421. https://doi.org/10.1007/s13355-013-0219-x Stork NE (2018) How many species of insects and other terrestrial arthropods are there on earth? Annu Rev Entomol 63(1):31–45. https://doi.org/10.1146/annurev-ento-020117-043348 Taberlet P, Coissac E, Pompanon F, Brochmann C, Willerslev E (2012) Towards next-generation biodiversity assessment using DNA metabarcoding. Mol Ecol 21(8):2045–2050. https://doi.org/10.1111/j.1365-294X.2012.05470.x Wickham H (2007) Reshaping data with the reshape package. Journal of Statistical Software 21(12):1 – 20. https://doi.org/10.18637/jss.v021.i12 Wickham H, Averick M, Bryan J, Chang W, McGowan LDA, François R, Grolemund G, Hayes A, Henry L, Hester J (2019) Welcome to the tidyverse. Journal of Open Source Software 4(43):1686. https://doi.org/10.21105/joss.01686 Young RG, Milián-García Y, Yu J, Bullas-Appleton E, Hanner RH (2021) Biosurveillance for invasive insect pest species using an environmental DNA metabarcoding approach and a high salt trap collection fluid. Ecol Evol 11(4):1558–1569. https://doi.org/10.1002/ece3.7113 Yu DW, Ji Y, Emerson BC, Wang X, Ye C, Yang C, Ding Z (2012) Biodiversity soup: metabarcoding of arthropods for rapid biodiversity assessment and biomonitoring. Methods Ecol Evol 3(4):613–623. https://doi.org/10.1111/j.2041-210X.2012.00198.x Zakaria N, Senawi J, Musa FH, Belabut D, Onn CK, Nor SM, Ahmad N (2014) Species composition of amphibians and reptiles in Krau Wildlife Reserve, Pahang, Peninsular Malaysia. Check List 10(2):335–343. https://doi.org/10.15560/10.2.335 Zar JH (2010) Biostatistical analysis, 5th Edition edn. Prentice Hall, Inc., Upper Saddle River, New Jersey Zenker MM, Specht A, Fonseca VG (2020) Assessing insect biodiversity with automatic light traps in Brazil: pearls and pitfalls of metabarcoding samples in preservative ethanol. Ecol Evol 10(5):2352–2366. https://doi.org/10.1002/ece3.6042 Zhang J, Kobert K, Flouri T, Stamatakis A (2014) PEAR: a fast and accurate Illumina paired-end read merger. Bioinformatics 30(5):614–620. http://dx.doi.org/10.1093/bioinformatics/btt593 Zizka VMA, Geiger MF, Hörren T, Kirse A, Noll NW, Schäffler L, Scherges AM, Sorg M (2022) Repeated subsamples during DNA extraction reveal increased diversity estimates in DNA metabarcoding of malaise traps. Ecol Evol 12(11):e9502. https://doi.org/10.1002/ece3.9502 Additional Declarations No competing interests reported. Supplementary Files SihalohoetalSI.pdf Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7367951","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":502901204,"identity":"bd7da107-bb13-47d1-b399-cc94075490ad","order_by":0,"name":"Hendra F. Sihaloho","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA9klEQVRIiWNgGAWjYBACAyCWADHYJBgYDzA2MMhBxNmI08IA0mLMw0asFgaolsQeQlrM2XsP3q6oYbDnk25+cODjDpv0/fI9Bgwfyg7j1GLZcy7Z8swxhsQ2mWMGB2eeScvtYeMxYJxxDrcWgxs5ZpINbAwJbBIJBod52w6DtTADGbi13H8D1PKPwZ5NIv0DSEs6D0jLX3xabvCYSTa2MTC2SeSAbUkAa2HEp+VMjrFlY59EIlBLwcGZbWmGPcfSCg72nEvHreX4GcObDd9s7OVnpG988LHNRp69+fDGBz/KrHFqgQIJVO4BQupHwSgYBaNgFOAHAG7jUt7J0qJUAAAAAElFTkSuQmCC","orcid":"","institution":"Texas Tech University","correspondingAuthor":true,"prefix":"","firstName":"Hendra","middleName":"F.","lastName":"Sihaloho","suffix":""},{"id":502901207,"identity":"68d4039a-4e15-41a4-8502-c27e63766aaa","order_by":1,"name":"Isham Azhar","email":"","orcid":"","institution":"Texas Tech University","correspondingAuthor":false,"prefix":"","firstName":"Isham","middleName":"","lastName":"Azhar","suffix":""},{"id":502901209,"identity":"f1a59920-8378-4d10-8499-c9733d587700","order_by":2,"name":"Millawati Gani","email":"","orcid":"","institution":"National Wildlife Forensic Laboratory, National Parks Peninsular Malaysia","correspondingAuthor":false,"prefix":"","firstName":"Millawati","middleName":"","lastName":"Gani","suffix":""},{"id":502901213,"identity":"b22797b3-e863-4924-8a7e-10739a588db8","order_by":3,"name":"Juliana Senawi","email":"","orcid":"","institution":"Universiti Kebangsaan Malaysia","correspondingAuthor":false,"prefix":"","firstName":"Juliana","middleName":"","lastName":"Senawi","suffix":""},{"id":502901216,"identity":"74c2471b-8d9a-404a-a2ef-af564559d0ab","order_by":4,"name":"Lim Shu Yong","email":"","orcid":"","institution":"Monash University Malaysia","correspondingAuthor":false,"prefix":"","firstName":"Lim","middleName":"Shu","lastName":"Yong","suffix":""},{"id":502901217,"identity":"524b69d1-1bb1-474c-a235-45cc3adf64c3","order_by":5,"name":"Qasim Ayub","email":"","orcid":"","institution":"Monash University Malaysia","correspondingAuthor":false,"prefix":"","firstName":"Qasim","middleName":"","lastName":"Ayub","suffix":""},{"id":502901219,"identity":"9cae6cb6-aa8e-4364-b321-961b05d6ff02","order_by":6,"name":"Rahul De","email":"","orcid":"","institution":"Texas Tech University","correspondingAuthor":false,"prefix":"","firstName":"Rahul","middleName":"","lastName":"De","suffix":""},{"id":502901220,"identity":"dfed964d-f9a9-4e11-a990-2a9a3dcb4718","order_by":7,"name":"Tigga Kingston","email":"","orcid":"","institution":"Texas Tech University","correspondingAuthor":false,"prefix":"","firstName":"Tigga","middleName":"","lastName":"Kingston","suffix":""},{"id":502901223,"identity":"b8b9c39a-c919-4d7b-99e5-0bd73a95b3a4","order_by":8,"name":"Caleb D. Phillips","email":"","orcid":"","institution":"Texas Tech University","correspondingAuthor":false,"prefix":"","firstName":"Caleb","middleName":"D.","lastName":"Phillips","suffix":""}],"badges":[],"createdAt":"2025-08-13 20:23:17","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7367951/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7367951/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":90569078,"identity":"5bf1afe9-908e-479a-ba28-75c3d46ab2d5","added_by":"auto","created_at":"2025-09-04 08:06:28","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":121302,"visible":true,"origin":"","legend":"\u003cp\u003eDatabase construction and bioinformatic workflow. Three databases were independently prepared\u003c/p\u003e\n\u003cp\u003ethrough the pipeline (left), and metabarcoding sequence data for each insect light trap subsample was summarized as 97% OTU similarity clusters (right).\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7367951/v1/8a96a96f966be6ac3c966456.png"},{"id":90569077,"identity":"d4d683ef-8fd5-4a1e-847d-a84858a23394","added_by":"auto","created_at":"2025-09-04 08:06:28","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":94222,"visible":true,"origin":"","legend":"\u003cp\u003eVenn diagrams illustrating taxonomic concordance of OTU assignments between blastn and tblastx. Best hits were filtered using the highest bitscore for each OTU (top row), while the bottom row used the highest bitscore and a percentage identity of ≥ 85% to obtain the best hits. Taxonomic concordances between blastn and tblastx were smaller at the lower taxonomic levels. For example, at the species (most right), using bitscore to get top hits, only 16.93% of OTUs (518 of 3,058) were assigned to the same species when queried using blastn and tblastx.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7367951/v1/a53676bed1497cf6e8e12754.png"},{"id":90569080,"identity":"3bd340ea-6210-429c-a108-b34fca429528","added_by":"auto","created_at":"2025-09-04 08:06:28","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":79789,"visible":true,"origin":"","legend":"\u003cp\u003eRarefaction and extrapolation based on number of subsamples for different Hill numbers (Hill\u003csub\u003e0\u003c/sub\u003e:\u003c/p\u003e\n\u003cp\u003erichness, Hill\u003csub\u003e1\u003c/sub\u003e: exponential Shannon-Wiener, Hill\u003csub\u003e2\u003c/sub\u003e: Inverse Simpson) and sample coverage curves for\u003c/p\u003e\n\u003cp\u003erichness (most right).\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7367951/v1/19ccc7b63f7743a22abf8a5e.png"},{"id":90570245,"identity":"66c8c63a-2193-460c-91c8-af6e7ba560d3","added_by":"auto","created_at":"2025-09-04 08:14:28","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":35965,"visible":true,"origin":"","legend":"\u003cp\u003eRarefaction and extrapolation based on number of trapping nights for different Hill numbers (Hill\u003csub\u003e0\u003c/sub\u003e:\u003c/p\u003e\n\u003cp\u003erichness, Hill\u003csub\u003e1\u003c/sub\u003e: exponential Shannon-Wiener, Hill\u003csub\u003e2\u003c/sub\u003e: Inverse Simpson) and their sample coverage curves for\u003c/p\u003e\n\u003cp\u003erichness (most right).\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7367951/v1/2cba8e513411ef4fd4cdbd64.png"},{"id":103282896,"identity":"a50c5527-ea30-4ffc-8d2d-d0e955248ccd","added_by":"auto","created_at":"2026-02-24 03:40:46","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":999930,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7367951/v1/1614c885-5e83-4f02-8556-928d9d69df3b.pdf"},{"id":90569085,"identity":"6a9ea91d-fd51-4e48-ae76-c5ac304cbeb5","added_by":"auto","created_at":"2025-09-04 08:06:28","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":264568,"visible":true,"origin":"","legend":"","description":"","filename":"SihalohoetalSI.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7367951/v1/f5a0a4041d1342324a388944.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Assessment of Insect Communities with Metabarcoding","fulltext":[{"header":"Introduction","content":"\u003cp\u003eInsects are the most speciose animal class, with a million described species and likely at least another 4.5\u0026nbsp;million (~\u0026thinsp;80%) that remain to be discovered (Stork \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Insects fill essential roles in ecosystems as pollinators, as well as in food webs, and consequently, insect diversity is a valuable measure of forest condition (Macedo-Reis et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; S\u0026aacute;nchez-Bayo and Wyckhuys \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). However, studying insect diversity can be challenging due to the vast number of species and individual numbers potentially encountered at a given site. Relatedly, morphology is often used to identify insect samples in surveys; however, this approach has a low throughput, being limited by the requirement for taxonomic expertise and the prevalence of undescribed species (O'Rourke et al. \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). The implementation of molecular approaches over the past few decades, namely, DNA barcoding (Hebert et al. \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2003a\u003c/span\u003e; Hebert et al. \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2003b\u003c/span\u003e; Ratnasingham and Hebert \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2007\u003c/span\u003e) metabarcoding techniques (Bik et al. \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Taberlet et al. \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Yu et al. \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2012\u003c/span\u003e), and the latter method has provided an alternative approach to insect identification that greatly increases throughput. Whereas genetic identification of individual specimens is commonly performed, metabarcoding refers to the sequencing of bulk samples or eDNA containing many individuals and taxa, followed by tabulation of the genetic diversity profiles (Liu et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Yu et al. \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Examples of sample types benefiting from metabarcoding include insect diversity surveys (Li et al. \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Roslin et al. \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Zenker et al. \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), pest surveillance (Batovska et al. \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Hardulak et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Young et al. \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) as well as investigations of insectivore diets by fecal analysis (Alberdi et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; O\u0026rsquo;Rourke et al. \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Phillips et al. \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2017\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eBulk samples from insect traps may be prepared for metabarcoding by homogenization prior to DNA extraction. However, traps commonly recover large masses of insects (e.g., from tens to hundreds of grams), but DNA extractions have a practical limit to the amount of input material, so it is not possible to extract from the entire homogenate. One solution is to reduce sample mass by dissecting out legs from individuals to carry forward through extraction (Ji et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Whereas this approach achieves the goal of reducing input mass, it may be impractical and time-consuming. An alternative approach is sample homogenization followed by subsampling (Zizka et al. \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Although a sample can theoretically be completely homogenized, in practice it seems unlikely that this will be achieved. Therefore, multiple subsamples are required to recover the diversity collected by the trap. A recent study using Malaise traps in a wetland reserve near Krefeld in Western Germany found that 12\u0026ndash;17 subsamples were required to cover 95% of the sample insect diversity (Zizka et al. \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). The homogenization approach employed by Zizka et al. (\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) represents a valuable contribution to insect metabarcoding techniques, and it is important to characterize the applicability of their recommendations to collections from different ecosystems, such as tropical forests, which have higher insect diversity than temperate regions (Gaston and Hudson \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e1994\u003c/span\u003e), and using different trapping methods, for instance, light traps that attract nocturnal flying insects (Shimoda and Honda \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2013\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eIn addition to sample processing considerations, the interpretation of results in a taxonomic framework based on comparison of identified sequences to a reference database is a common component of analysis and inference (Magoga et al. \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). A few technical issues potentially limit accurate taxonomic assignment. First, despite great effort (i.e., Barcode of Life Database (BOLD); Ratnasingham and Hebert \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2007\u003c/span\u003e), representative DNA sequences for most of the world\u0026rsquo;s insect taxa are not currently available, and will not be so for the foreseeable future. Second, due to sequencing and bioinformatic technical constraints, metabarcoding approaches have relied on relatively short mitogenomic region as the genetic barcode locus. Thus, database preparation and taxonomic interpretation should carefully consider taxonomic reliability.\u003c/p\u003e\u003cp\u003eWe sampled insects using light traps on the Tengku Hassanal Wildlife Reserve (THWR) in Pahang, Malaysia. The THWR, recently renamed from Krau Wildlife Reserve, is the third largest protected area in Peninsular Malaysia (Zakaria et al. \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Like other areas in Southeast Asia, the area around THWR has been extensively converted to non-timber plantations (rubber \u003cem\u003eHevea brasiliensis\u003c/em\u003e, oil palm \u003cem\u003eElaeis guineensis\u003c/em\u003e), resulting in a mosaic landscape of various land uses and forest fragments. The goal of this work was to contribute to the body of best practice for characterizing insect diversity by bulk sampling and metabarcoding, specifically by sampling in a tropical rainforest with light traps. We sought to estimate: (1) the number of subsamples from homogenates required to recover at least 90% sample diversity; (2) the number of sampling nights required to detect 90% of insect OTU diversity at a locality over time; (3) evaluate the best approach and limitations of taxonomic inference.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eStudy Sites\u003c/h2\u003e\u003cp\u003eSampling was conducted from May 26th to July 3rd, 2019, at Tengku Hassanal Wildlife Reserve in Pahang, Malaysia (THWR; Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e; 50m asl). Those three months were previously reported to have peak insect biomass and rainfall in 2009 and 2010 (Nurul-Ain et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). The reserve covers an area of 60,349 hectares of continuous dipterocarp forest under the management of the Department of Wildlife and National Parks (PERHILITAN). The average annual rainfall of the area is about 2,000 mm, and the daily temperature ranges from 23\u003csup\u003eo\u003c/sup\u003eC to 33\u003csup\u003eo\u003c/sup\u003eC (Zakaria et al. \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Insect trapping was conducted for six nights at six different locations (Supp.Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) at Kuala Lompat Research Station, which is located in the southeastern part of the reserve.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eLight trap samples: Sample ID, Subsample ID, Sampling dates, GPS coordinates, and the wet weight.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"6\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSample ID\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSubsample ID\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eSampling Date\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eLatitude (\u003csup\u003eo\u003c/sup\u003eN)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eLongitude (\u003csup\u003eo\u003c/sup\u003eE)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eWet weight (g)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTK05\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eTK05A - TK05L\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e03 Jul 2019\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e3.7134\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e102.2739\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e165.1\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTK06\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eTK06A - TK06L\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e20 Jun 2019\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e3.7177\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e102.2763\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e137.9\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTK07\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eTK07A - TK07L\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e22 Jun 2019\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e3.7182\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e102.2737\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e445.4\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTK08\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eTK08A - TK08L\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e06 Jun 2019\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e3.7150\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e102.2786\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e197.9\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTK09\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eTK09A - TK09L\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e27 Jun 2019\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e3.7157\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e102.2759\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e185.1\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTK10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eTK10A - TK10L\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e26 May 2019\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e3.7151\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e102.2821\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e157.6\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003emean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e214.83\u0026thinsp;\u0026plusmn;\u0026thinsp;114.89\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eTrapping Methods\u003c/h3\u003e\n\u003cp\u003eInsects were collected using light traps (Leptraps LLC) deployed across a 1 km\u003csup\u003e2\u003c/sup\u003e area of THWR, and the median distance between traps was 531 m. Traps were equipped with an 18-inch 12V 32-watt quantum black light bulb with rigid plexiglass vanes, a photoelectric switch, and a funnel placed at the opening of a plastic bucket containing Propylene Glycol (PG) (Moreau et al. \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Patrick et al. \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2016\u003c/span\u003e) for preserving collected insects. Light traps were powered by 12V 20AH sealed lead acid batteries. They were deployed daily around 6.30pm (dusk) and collected around 6.30am (dawn) the following morning. Insect samples from each trap were poured into a collection jar labelled with trap number, date, and time of collection. The insect jars were kept at ambient temperature before being processed in the laboratory.\u003c/p\u003e\n\u003ch3\u003eSample Processing\u003c/h3\u003e\n\u003cp\u003eContents of each insect jar were processed separately, starting by pouring insects onto a 0.8 mm wire mesh stainless steel sieve (Hallmann et al. \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Sieving continued until drainage droplets were separated by 10 seconds or more. The sieved sample was then weighed using a scale with an accuracy of at least 0.1 g to obtain bulk biomass, which we refer to hereafter as wet weight. The sample was transferred to a cleaned blender (Model 7010S, Waring Commercial) for homogenization. The homogenization was carried out by blending the insect samples with 50 mL digestion buffer (pH 8.0) for 90 seconds. Fresh digestion buffer was prepared before processing insect samples using 10 mM Tris-HCL, 10 mM NaCl, 5 mM CaCl\u003csub\u003e2\u003c/sub\u003e, 2.5 mM EDTA, and 2\u0026permil; SDS (Campos and Gilbert \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Following the homogenization, 25 mL of the homogenate was transferred into 50 mL falcon tubes and stored at -80\u003csup\u003eo\u003c/sup\u003eC. Each homogenate was subsampled 12 times by thawing the samples in an ice bucket, mixing thawed homogenate using a vortex, then using a blunted 1 mL pipet tip to pipet 100 \u0026micro;L of homogenate into 2 mL sterile vials containing 1 mL of fresh digestion buffer, this was the same digestion buffer previously described with additional 50mM DTT and 100 \u0026micro;g proteinase K. Homogenate was re-vortexed between each 100 \u0026micro;L subsample. The subsample vials were incubated in 50\u003csup\u003eo\u003c/sup\u003eC water bath for at least 18 hours before they were stored at -80\u003csup\u003eo\u003c/sup\u003eC. DNA was extracted from each subsample using DNeasy Blood \u0026amp; Tissue Kit (Qiagen Inc.) following the manufacturer\u0026rsquo;s protocol. Each DNA extract was assessed for quality and quantity using Nanodrop (Thermo Fisher Scientific Inc.).\u003c/p\u003e\n\u003ch3\u003eCytochrome Oxidase I Amplification and Sequencing\u003c/h3\u003e\n\u003cp\u003eMetabarcoding was employed to assign the taxonomy of insect Operational Taxonomic Units (OTU). A region of Cytochrome Oxidase I (COI) was amplified using mlCOIintF (Leray et al. \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2013\u003c/span\u003e) and HCO2198R (Folmer et al. \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e1994\u003c/span\u003e) primers that targeted a 313 bp amplicon length. Both primers were designed to include the Illumina 5' overhang, so that the Illumina barcodes could be directly added to the purified PCR products in the second round of amplification. The forward and reverse primers used in the first round PCR were TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG\u003cb\u003eGGWACWGGWTGAACWGTWTAYCCYCC\u003c/b\u003e and GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG\u003cb\u003eTAAACTTCAGGGTGACCAAAAAATCA\u003c/b\u003e (the sequences of mlCOIintF and HCO2198R primers, respectively, are boldface). The first round of PCR reactions was carried out in 25 \u0026micro;L total volumes containing 2.5 \u0026micro;L DNA template (~\u0026thinsp;3ng), 1X Q5 Hifi Hotstart ReadyMix (New England Biolabs Inc.), and 0.2 \u0026micro;M of each primer. The following temperature profile was used to amplify the COI region: A 5min initial denaturation at 98\u003csup\u003eo\u003c/sup\u003eC, followed by 25 cycles of 30s denaturation at 98\u003csup\u003eo\u003c/sup\u003eC, 40s annealing at 54\u003csup\u003eo\u003c/sup\u003eC, 1min extension at 72\u003csup\u003eo\u003c/sup\u003eC, and a final extension at 72\u003csup\u003eo\u003c/sup\u003eC for 10min. PCR products were purified using 0.8X AMPure XP beads (Beckman Coulter, Inc., IN, USA). The second round of amplifications was performed to incorporate the Illumina i5 and i7 adapters and 8bp barcodes. These reactions were performed in a volume of 10 \u0026micro;L, containing 5 \u0026micro;L of KAPA HiFi HotStart Ready Mix (Roche Diagnostics), 1 \u0026micro;L each of i5 and i7 barcoded adapters at 5 pM, and 3 \u0026micro;L of purified PCR products. The thermal profile for the second PCR was 3min initial denaturation at 95\u003csup\u003eo\u003c/sup\u003eC, followed by 8 cycles of 30s denaturation at 95\u003csup\u003eo\u003c/sup\u003eC, 30s annealing at 55\u003csup\u003eo\u003c/sup\u003eC, 30s extension at 72\u003csup\u003eo\u003c/sup\u003eC, and a final extension at 72\u003csup\u003eo\u003c/sup\u003eC for 1min. Libraries were purified using 0.7X AMPure XP beads (Beckman Coulter, Inc., IN, USA). Then, libraries were analyzed using a Qubit fluorometer (Qubit dsDNA HS Assay Kit; Life Technologies, CA, USA) and TapeStation (High Sensitivity D1000 ScreenTape assay; Agilent Technologies, Inc., CA, USA) to determine concentrations and fragment sizes, respectively, and subsequently pooled in equimolar amounts. The pooled libraries were sequenced on a MiSeq platform using 2 x 300 bp sequencing and MiSeq Reagent Kit v3 (Illumina, Inc.) at the Monash University Malaysia Genomics Platform.\u003c/p\u003e\n\u003ch3\u003eBuilding the insect COI reference database\u003c/h3\u003e\n\u003cp\u003eWe developed a database from the BOLD repository (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Insect sequences were downloaded from BOLD (Ratnasingham and Hebert \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2007\u003c/span\u003e) in December 2022 using modified R scripts by O'Rourke et al. (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) in RStudio (Posit team \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; R Core Team \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Our database pipeline was modified from previously described approaches (O'Rourke et al. \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Robeson et al. \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) and included removal of any leading and trailing Ns (uncalled bases), followed by removal of homopolymers equal to or greater than 12 base pairs and ambiguous nucleotide runs equal to or greater than 6 base pairs. Obtaining good starting alignment of millions of sequences is computationally challenging, so we first reduced the database to Malaysian entries, which were subset to 655\u0026ndash;1500 bp length entries and aligned using MAFFT v7.490 (Katoh and Standley \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). The mlCOIintF and HCO2198R primers (Folmer et al. \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e1994\u003c/span\u003e; Leray et al. \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2013\u003c/span\u003e) were then added to this alignment to verify coordinates and the length of the targeted sequencing region. Next, the remaining database entries were added to the alignment using MAFFT v7.490 (Katoh and Standley \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2013\u003c/span\u003e) with \u0026ndash;auto \u0026ndash;addfull flags. The aligned region of the database was extracted using extract_alignment_region.py script (Robeson et al. \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), and sequences were then trimmed to the COI region flanked by primers described above (300\u0026ndash;319 bp) using SeqKit v2.3.0 (Shen et al. \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). The Nuclear mitochondrial DNA segments (NUMTs) were then removed using the Biostrings v2.72.1 package in R (Pag\u0026egrave;s et al. \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Next, the unique sequences were identified with Usearch v11 --fastx_uniques command (Edgar \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). Any unique sequences represented by multiple sequences with conflicting taxonomy were removed.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eProcessing of COI metabarcoding data\u003c/h2\u003e\u003cp\u003eMetabarcoding sequence data were processed following Phillips et al. (\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Briefly, primers were removed from raw reads using cutadapt v2.6 (Martin \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). Paired-end reads were merged using PEAR v0.9.11 with default settings (Zhang et al. \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Usearch v11 (Edgar \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2010\u003c/span\u003e) was employed for quality filtering, identifying and abundance counting of unique sequences, and clustering based on 97% similarity to inform OTU table construction. A multiple sequence alignment of the representative OTU sequences was created by first using MAFFT v7.490 (Katoh and Standley \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2013\u003c/span\u003e) with the \u0026ndash;auto flag applied to a subsample of 200 sequences. Next, the remaining representative OTU sequences were added and aligned with \u0026ndash;auto \u0026ndash;addfull flags. Any OTUs with a length less than 300 were then removed using SeqKit v2.3.0 (Shen et al. \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). The OTUs containing NUMTs were removed using Biostrings v2.72.1 R package (Pag\u0026egrave;s et al. \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Data management was carried out using the phyloseq v1.48 R package (McMurdie and Holmes \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2013\u003c/span\u003e) in RStudio (Posit team \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; R Core Team \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Taxonomy for each OTU was independently assigned using blastn and tblastx (v2.14.1+; Camacho et al. \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2009\u003c/span\u003e) with an e-value threshold of 1e-4. For each OTU, we retained the top hit based on bit score value. Both blastn and tblastx were evaluated to understand if nucleotide- versus protein-level comparison offered advantages. The top bit score, which is a widely used measure of alignment significance, was recorded for each OTU for each database. Bit score was selected as the comparator metric because it is comparable across databases of different sizes and more reliable than percent identity for inferring homology (Pearson \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Since the values of bitscore from blastn and tblastx had different possible ranges, each was transformed using the proportion of maximum (POM) before comparison using the Wilcoxon signed rank test (Quinn and Keough \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2002\u003c/span\u003e; Zar \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2010\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eBiological diversity and sample coverage\u003c/h3\u003e\n\u003cp\u003eHill numbers (0, 1, 2) were calculated to summarize insect OTU diversity detected from subsamples and for entire light traps (Alberdi and Gilbert \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Chao et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Unbiased comparison of communities requires parity of sample coverage, defined as \u0026ldquo;\u003cem\u003ethe proportion of the total number of individuals in a community that belong to the species represented in the sample\u003c/em\u003e (Chao and Jost \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Sample coverage estimates how well true diversity is obtained by a sample (Roswell et al. \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). One way to approach parity is to set a coverage threshold and then compare communities at this threshold. In this study, we wanted to determine the number of subsamples that would recover at least 90% and 95% of insect OTU diversity for each sample. Sample coverage was estimated at two levels: (1) the number of subsamples per light trap homogenate required to recover insect OTU diversity for each light trap, and (2) the number of sampling nights needed to characterize insect OTU diversity at a locality. Sample coverage was calculated using the iNEXT v2.0.20 R package (Chao et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Hsieh et al. \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). The OTU table was normalized to the lowest sample-wide sequencing depth using stratified random sampling (Beule and Karlovsky \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Then, sample coverage was calculated to determine the number of subsamples required to represent at least 90% and 95% coverage for each sample. Each light trap sample was subsampled 12 times, and extrapolation was limited to 24 subsamples to avoid estimation bias (Chao et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Chao and Jost \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Also, to evaluate the efficacy of homogenization, one-way Analysis of Variance (ANOVA) was used to test if subsample combination affected diversity recovered (subsample combination at a given subsampling depth was a nested variable within light trap). At the locality-level, sample coverage was calculated to determine the number of light traps required to represent at least 90% and 95% of local diversity. Analyses in R also employed the following packages: bold v1.3.0 (Dubois and Chamberlain \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), SRS v0.23 (Beule and Karlovsky \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Heidrich et al. \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), taxize v0.9.100 (Chamberlain and Sz\u0026ouml;cs \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Chamberlain et al. \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), tidyverse v2.0.0 (Wickham et al. \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), VennDiagram v1.7.3 (Chen \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), and reshape2 (Wickham \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2007\u003c/span\u003e).\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003eInsect OTU diversity\u003c/h2\u003e\u003cp\u003eA total of 3,329,140 paired-end reads were obtained study-wide with 23,119\u0026thinsp;\u0026plusmn;\u0026thinsp;2,611 (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;sd) per subsample. The number of reads retained through bioinformatic processing across subsamples ranged from 15,788 to 28,222, with an average of 20,615\u0026thinsp;\u0026plusmn;\u0026thinsp;2,524. The range in wet weight from the six light traps was 137.9 to 445.4 g (average 214.83\u0026thinsp;\u0026plusmn;\u0026thinsp;114.89 g; Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), from which 3,058 OTUs at the 97% similarity level were discovered. The average number of insect OTUs per trap subsample was 508\u0026thinsp;\u0026plusmn;\u0026thinsp;136. An analysis of variance (ANOVA) was conducted to assess the efficacy of homogenization, and it indicated no significant effect of subsample combination on insect OTU richness (p\u0026thinsp;\u0026gt;\u0026thinsp;0.05; Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003cem\u003eReference database\u003c/em\u003e\u003c/p\u003e\u003cp\u003eA total of 4,241,641 database entries were downloaded from BOLD using \u0026ldquo;insecta\u0026rdquo; as search query criteria and filtered for marker codes \u0026ldquo;COI-5P\u0026rdquo;. Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e summarizes the number of starting and ending database entries retained at each step in the database development workflow. It showed that the final database only preserved 17.14% of the original database entries (726,899 sequences).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eNumber of candidate database sequence entries (class insecta) retained at each step of the database development workflow.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"3\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDevelopment Steps\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNumber of entries\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eRetain (%)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRaw entries\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e4,241,641\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e100.00\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRemoving homopolymer\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e4,229,603\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e99.72\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTrimming to the sequenced region\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e2,484,568\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e58.58\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRemove NUMTs\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e2,453,660\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e57.85\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRetain unique sequences\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e745,041\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e17.56\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRemove taxonomic discrepancies\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e726,899\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e17.14\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eFor the total 3,058 insect OTUs, blastn analysis returned significant hits for 3,056 OTUs, while tblastx analysis provided significant matches for all OTUs for each database. The median nucleotide (blastn) and translated protein (tblastx) match identities were 92.01% and 93.20%. Comparison of standardized bit scores between blastn and tblastx revealed that tblastx was significantly better (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), but their percent identity distributions did not significantly differ (p\u0026thinsp;\u0026gt;\u0026thinsp;0.05). The number of insect OTUs with significant matches to either blastn or tblastx searches varied across taxonomic levels (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). In general, the concordance of taxonomic assignment between blastn and tblastx was smaller at lower taxonomic levels. Out of 3,058 insect OTUs, only 518 (16.94%) had concordant species-level identifications between blastn and tblastx. This number increased to 878 OTUs (28.71%) at the generic level, 2,045 (66.87%) at the familial level, and 2,858 (93.46%) at the ordinal level. A similar trend was also observed when a minimum 85% identity parameter was applied to the best blastn or tblastx hits (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, bottom row). Although it was not significant, the number of OTUs with concordant taxonomic assignments decreased when an 85% PID was applied.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003eLight Trap Diversity and Completeness\u003c/h2\u003e\u003cp\u003eThe diversity of insect OTUs contained in each of the 12 subsamples for each of six light traps was summarized by a Hill\u003csub\u003e0\u003c/sub\u003e (richness) of 902.5\u0026thinsp;\u0026plusmn;\u0026thinsp;243.05, Hill\u003csub\u003e1\u003c/sub\u003e (exponential Shannon-Wiener) 689.59\u0026thinsp;\u0026plusmn;\u0026thinsp;199.44, and Hill\u003csub\u003e2\u003c/sub\u003e (Inverse Simpson) 625.27\u0026thinsp;\u0026plusmn;\u0026thinsp;183.01. As expected, increasing the number of subsamples for a given light trap enhanced the estimated sample completeness (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e4\u003c/span\u003e). All samples, except TK06, reached 98% sample coverage when using 12 subsamples (98.13\u0026thinsp;\u0026plusmn;\u0026thinsp;1.00); light trap TK06 only reached 96.2% sample coverage when using 12 subsamples, but TK06 was not an outlier with respect to wet weight (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). With the intent to identify the minimum number of required subsamples to detect at least 90% insect OTUs diversity, four subsamples per light trap achieved this threshold, averaging 93.83% \u0026plusmn; 1.62. For comparison, a minimum of 8 subsamples was required to detect at least 95% insect OTU diversity in a light trap.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe six insect light traps studied here came from six different nights spanning 39 days (May 26th to July 3rd, 2019) within a forest habitat\u0026thinsp;\u0026lt;\u0026thinsp;1 km\u003csup\u003e2\u003c/sup\u003e. Using four subsamples per light trap, the number of traps needed to discover insect OTU diversity in the given spatiotemporal setting was investigated. The diversity of insect OTU for six nights was 2,701 (Hill\u003csub\u003e0\u003c/sub\u003e), 2,267 (Hill\u003csub\u003e1\u003c/sub\u003e), and 1,845 (Hill\u003csub\u003e2\u003c/sub\u003e). Results indicated that six sampling nights only covered 59.8% of insect OTU diversity (Fig.\u0026nbsp;5). Extrapolation, which has estimation uncertainty due to needing to extrapolate greater than double the number of actual sample nights, indicated that 21 and 28 trap nights would be required to reach 90% and 95% OTU diversity of nocturnal flying insects, respectively.\u003c/p\u003e\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eStudies of insect diversity generally follow two approaches: non-destructive and destructive sampling. The non-destructive methods, such as morphological identification, ethanol-based DNA extraction of parts of individual specimens, and imaging prior to DNA extraction, offer the key advantage of preserving physical specimens for future examination, taxonomic validation, or voucher deposition. However, these methods are often more labor-intensive and may not scale efficiently for large biodiversity assessments. In contrast, destructive metabarcoding approaches, such as the homogenization of bulk samples (Yu et al. \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2012\u003c/span\u003e), prioritize high-throughput processing and standardization. In our current study, we adopted the homogenization approach with the goal of optimizing the number of subsamples required to effectively capture the majority of molecular operational taxonomic units (OTUs) within a bulk insect sample. While this method sacrifices the preservation of individual specimens, it provides a practical and efficient solution for broad-scale biodiversity assessment.\u003c/p\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003eInsect diversity and reference database\u003c/h2\u003e\u003cp\u003eA total of 3,058 97% insect OTUs were identified in this study from 6 light traps collected on separate nights across a 39-day span within a 1 km\u003csup\u003e2\u003c/sup\u003e forest area in peninsular Malaysia. Total diversity reported herein is higher than that reported in a study of insect pests in diverse biomes of Brazil (Zenker et al. \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) as well as one from a German wetland (Zizka et al. \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Many differences among studies, including sampling duration, geographic spread, trapping method, molecular and analytical choices, preclude direct comparisons. However, given both the large volume of insects attracted by the light traps used herein, and the high biodiversity of the study system, the provided recommendations on sample treatment are likely to be applicable across many different study systems. Also, there is no reason to expect that the challenges encountered with taxonomic assignment are unique to this study.\u003c/p\u003e\u003cp\u003eThere is an estimated high percentage of insect biodiversity for which genetic representatives are lacking. Given the disparity between what is known and what will be encountered in a metabarcoding study at a given site, it seems intuitive that comparing observed genetic lineages to all available information is the best option for taxonomic assignment. Still, database preparation herein revealed that for the COI gene region considered, there were instances of multiple BOLD entries having identical sequences (~\u0026thinsp;1.7\u0026nbsp;million) as well as entries excluding the region targeted by the assay (~\u0026thinsp;1.7\u0026nbsp;million). In total, about 3.4\u0026nbsp;million sequences, equating to about 40% of sequences considered for database inclusion, were identified on these criteria. Representing such sequence duplicates as a single representative sequence with taxonomy trimmed back to the lowest congruent level serves to both reduce database size and reflect taxonomic uncertainty.\u003c/p\u003e\u003cp\u003eInsect OTUs were compared to the reference database using both nucleotide (blastn) and protein (tblastx) based approaches, with the rationale being to assess if queries leveraging more (nucleotide) or less (protein) variable sequences offered any advantages. From direct comparison, relatively high taxonomic concordance was observed at the ordinal level, however, low concordances elsewhere (familial (68%), genera (41%), and species (29%) levels). Also, average percentage identities for nucleotide- and protein-level comparisons did not significantly differ and were, at around 92\u0026ndash;93%, below that commonly expected for intraspecific comparisons. The general taxonomic uncertainty is thought to be caused not only by a lack of sequence representatives for much Insecta diversity, but also due to a lack of species-level resolution influenced by homoplasy in a small genomic region.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003eSample Completeness\u003c/h2\u003e\u003cp\u003eA main goal herein was to understand effort requirements to capture 90 or 95% insect OTU diversity in homogenate samples. Therefore, subsequent investigations into requirements for light trap subsampling and field efforts were conducted using 97% OTUs as the level of evolutionary delineation. Rarefaction curves (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e3\u003c/span\u003e) suggested that additional effort beyond four subsamples would, on average, only cover an additional 6% of diversity disproportionately comprised of rare OTUs. For instance, for obtaining 95% sample coverage, a total of eight subsamples was required. Whereas in this study, each subsample was amplified and sequenced separately to allow investigation of the subsampling procedure, future work following these guidelines can pool subsamples\u0026rsquo; DNA after extraction and prior to sequencing for cost savings. Thus, achieving 95% coverage would not require much expense beyond that required for 90% sample coverage. That said, pooling subsamples could introduce PCR biases, so treating subsamples independently throughout the process is preferred if cost is not prohibitive.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eOther recent work by Zizka et al. (\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) was structured similarly to the current study but used Malaise traps and conducted field work in a wetland nature reserve near Krefeld in western Germany. That study reported that 12\u0026ndash;17 subsamples were required to capture 95% diversity, a level of diversity detected by about eight subsamples in the current study. This is a surprising contrast, especially given that tropical rainforests support more insect diversity than temperate regions (Lewinsohn and Roslin \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2008\u003c/span\u003e). Multiple aspects of their study design and analytical approach differed from those reported herein, which could explain this difference. One possible key difference between the two studies is the amount of sequencing effort per subsample; the current study had an average of 23,119 reads per subsample, whereas Zizka et al. (\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) reported an average of 154,659 reads per subsample. It is simultaneously possible that the 7-fold increased sequencing effort by Zizka et al. (\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) allowed greater detection of rare OTUs, but also lead to the creation of erroneous OTUs through recurrent sequencing error across great sequencing depth. The latter is a real possibility given Illumina\u0026rsquo;s average error rate of 1 out of 1000 base calls in conjunction with deep sequencing of PCR amplicons. It is not possible to delineate these effects, but re-analysis of the Zizka et al. (\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) data set using the bioinformatic approach reported herein with sequence data first subsampled to the current studies\u0026rsquo; average sequencing effort also resulted in four subsamples detecting at least 90% (93.82\u0026thinsp;\u0026plusmn;\u0026thinsp;0.99) total diversity across all samples (Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e, Table S2). Thus, a reasonable joint interpretation appears to be that four subsamples is a generalizable recommendation to achieve a minimum 90% sample coverage so long as sequencing effort is regulated, and additional subsampling will provide more but modest additional information.\u003c/p\u003e\u003cp\u003eFor purposes of sample coverage estimation, extrapolation greater than two times the number of actual samples has been cautioned because of increased bias (Chao et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Nevertheless, to provide some insight into how much trapping effort would be required to summarize insect OTU diversity at the study site adequately, it was estimated that 21 sampling nights, with one trap set per night, would be required to reach 90% completeness. This is with respect to the 39 day and 1 km\u003csup\u003e2\u003c/sup\u003e spatiotemporal interval from which the 6 samples were obtained. Our interpretation is that insect biodiversity turnover is quite high in the Malaysian rainforest study site, and with respect to the spatial scale of the current study, trapping would need to occur about every other night. This requirement would obviously change depending on spatiotemporal scale. Interesting future work possibilities include longitudinal monitoring of numerous set trapping stations across the landscape.\u003c/p\u003e\u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eOptimizing resources (funds, time, personnel) is crucial to the design of ecological fieldwork, especially in rapid biodiversity assessments, such as monitoring or large-scale biodiversity projects. Results demonstrate the feasibility of the reported light trap homogenization strategy and suggest that four to eight subsamples of homogenate should detect the large majority of insect OTU diversity in bulk collections. Contrasts with other similar work suggest that limiting sequencing depth is an important consideration to avoid the creation of false OTUs due to excessive sequencing effort. Finally, given the high insect diversity in tropical forests, trapping for approximately 20 nights within a roughly 1 km\u0026sup2; area over a one-month time frame is expected to facilitate the detection of the large majority of nocturnal flying insect diversity. Studies moving forward should verify subsampling species accumulation wherever possible, but we expect these suggestions to be generally applicable.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eAcknowledgement\u003c/p\u003e\n\u003cp\u003eThe authors thank the Economic Planning Unit of the Prime Minister\u0026rsquo;s Department (EPU 40/200/19/3510), \u0026nbsp;Department of Wildlife and National Parks (PERHILITAN) Peninsular Malaysia\u0026nbsp;(JPHL \u0026amp; TN (IP): 100-34/1.24 Jld. 14 (31),\u0026nbsp;and Access \u0026amp; Benefit Sharing - Pahang Biodiversity Council (MBP.600(S)-1/1/1 Jld. 4 (10) for permission to conduct research in Malaysia. We\u0026nbsp;thank the Natural Science Research Laboratory, Museum of Texas Tech University, for assistance with field supplies and organization, and Sze Mei Lee for her technical assistance in preparing some of the DNA libraries.\u0026nbsp;This material is based upon work supported by the National Science Foundation under Grant No. DEB-1754810. LSY and QA were supported by a strategic grant from the School of Science, Monash University Malaysia (Grant No I-M010-STG-000188).\u003c/p\u003e\n\u003cp\u003eConflict of Interest\u003c/p\u003e\n\u003cp\u003eThe authors declare no conflict of interest.\u003c/p\u003e\n\u003cp\u003eAuthors\u0026apos; Contributions\u003c/p\u003e\n\u003cp\u003eThis paper was completed through collaboration among all authors. TK and CDP designed and acquired funding for the project. IA, TK, JS and CDP carried out the fieldwork. MG, LSY, and QA conducted laboratory procedures. HFS, CDP, IA, TK and RD analyzed the data. HFS \u0026amp; CDP wrote the first draft of the manuscript. All authors reviewed the manuscript and provided critical advice.\u003c/p\u003e\n\u003cp\u003eData Availability Statement\u003c/p\u003e\n\u003cp\u003eThe raw sequence data generated in this study are available at GenBank Sequence Read Archive, BioProject accession number PRJNA1260500. All scripts of the current study can be accessed at https://github.com/mhenso/insect_metabarcoding.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eAlberdi A, Gilbert MTP (2019) A guide to the application of Hill numbers to DNA-based diversity analyses. Mol Ecol Resour 19(4):804\u0026ndash;817. https://doi.org/10.1111/1755-0998.13014\u003c/li\u003e\n \u003cli\u003eAlberdi A, Razgour O, Aizpurua O, Novella-Fernandez R, Aihartza J, Budinski I, Garin I, Ib\u0026aacute;\u0026ntilde;ez C, Izagirre E, Rebelo H, Russo D, Vlaschenko A, Zhelyazkova V, Zrnčić V, Gilbert MTP (2020) DNA metabarcoding and spatial modelling link diet diversification with distribution homogeneity in European bats. Nat Commun 11(1):1154. https://doi.org/10.1038/s41467-020-14961-2\u003c/li\u003e\n \u003cli\u003eBatovska J, Piper AM, Valenzuela I, Cunningham JP, Blacket MJ (2021) Developing a non-destructive metabarcoding protocol for detection of pest insects in bulk trap catches. Sci Rep 11(1):7946. https://doi.org/10.1038/s41598-021-85855-6\u003c/li\u003e\n \u003cli\u003eBeule L, Karlovsky P (2020) Improved normalization of species count data in ecology by scaling with ranked subsampling (SRS): application to microbial communities. PeerJ 8:e9593. https://doi.org/10.7717/peerj.9593\u003c/li\u003e\n \u003cli\u003eBik HM, Porazinska DL, Creer S, Caporaso JG, Knight R, Thomas WK (2012) Sequencing our way towards understanding global eukaryotic biodiversity. Trends Ecol Evol 27(4):233\u0026ndash;243. https://doi.org/10.1016/j.tree.2011.11.010\u003c/li\u003e\n \u003cli\u003eCamacho C, Coulouris G, Avagyan V, Ma N, Papadopoulos J, Bealer K, Madden TL (2009) BLAST+: architecture and applications. BMC Bioinformatics 10(1):421. https://doi.org/10.1186/1471-2105-10-421\u003c/li\u003e\n \u003cli\u003eCampos PF, Gilbert TMP (2012) DNA extraction from keratin and chitin. In: Shapiro B, Hofreiter M (eds) Ancient DNA: methods and protocols. Humana Press, Totowa, NJ, pp 43\u0026ndash;49\u003c/li\u003e\n \u003cli\u003eChamberlain SA, Sz\u0026ouml;cs E (2013) taxize: taxonomic search and retrieval in R. F1000Research 2(191). https://doi.org/10.12688/f1000research.2-191.v2\u003c/li\u003e\n \u003cli\u003eChamberlain SA, Szoecs E, Foster Z, Arendsee Z, Boettiger C, Ram K, Bartomeus I, Baumgartner J, O\u0026apos;Donnell J, Oksanen J, Tzovaras BG, Marchand P, Tran V, Salmon M, Li G, Greni\u0026eacute; M (2020) taxize: Taxonomic information from around the web\u003c/li\u003e\n \u003cli\u003eChao A, Gotelli NJ, Hsieh TC, Sander EL, Ma KH, Colwell RK, Ellison AM (2014) Rarefaction and extrapolation with Hill numbers: a framework for sampling and estimation in species diversity studies. Ecol Monogr 84(1):45\u0026ndash;67. https://doi.org/10.1890/13-0133.1\u003c/li\u003e\n \u003cli\u003eChao A, Jost L (2012) Coverage-based rarefaction and extrapolation: standardizing samples by completeness rather than size. Ecology 93(12):2533\u0026ndash;2547. https://doi.org/10.1890/11-1952.1\u003c/li\u003e\n \u003cli\u003eChen H (2022) VennDiagram: generate high-resolution Venn and Euler plots\u003c/li\u003e\n \u003cli\u003eDubois S, Chamberlain S (2023) bold: Interface to Bold Systems API\u003c/li\u003e\n \u003cli\u003eEdgar RC (2010) Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26(19):2460\u0026ndash;2461. https://doi.org/10.1093/bioinformatics/btq461\u003c/li\u003e\n \u003cli\u003eFolmer O, Black M, Hoeh W, Lutz R, Vrijenhoek R (1994) DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Mol Mar Biol Biotechnol 3(5):294\u0026ndash;299\u003c/li\u003e\n \u003cli\u003eGaston KJ, Hudson E (1994) Regional patterns of diversity and estimates of global insect species richness. Biodivers Conserv 3(6):493\u0026ndash;500. https://doi.org/10.1007/BF00115155\u003c/li\u003e\n \u003cli\u003eHallmann CA, Sorg M, Jongejans E, Siepel H, Hofland N, Schwan H, Stenmans W, M\u0026uuml;ller A, Sumser H, H\u0026ouml;rren T, Goulson D, de Kroon H (2017) More than 75 percent decline over 27 years in total flying insect biomass in protected areas. PLoS One 12(10):e0185809. https://doi.org/10.1371/journal.pone.0185809\u003c/li\u003e\n \u003cli\u003eHardulak LA, Morini\u0026egrave;re J, Hausmann A, Hendrich L, Schmidt S, Doczkal D, M\u0026uuml;ller J, Hebert PDN, Haszprunar G (2020) DNA metabarcoding for biodiversity monitoring in a national park: screening for invasive and pest species. Mol Ecol Resour 20(6):1542\u0026ndash;1557. https://doi.org/10.1111/1755-0998.13212\u003c/li\u003e\n \u003cli\u003eHebert PDN, Cywinska A, Ball SL, deWaard JR (2003a) Biological identifications through DNA barcodes. Proc R Soc Lond B Biol Sci 270(1512):313\u0026ndash;321. 10.1098/rspb.2002.2218\u003c/li\u003e\n \u003cli\u003eHebert PDN, Ratnasingham S, de Waard JR (2003b) Barcoding animal life: cytochrome \u003cem\u003ec\u003c/em\u003e oxidase subunit 1 divergences among closely related species. Proc R Soc Lond B Biol Sci 270(Suppl 1):S96\u0026ndash;S99. https://doi.org/10.1098/rsbl.2003.0025\u003c/li\u003e\n \u003cli\u003eHeidrich V, Karlovsky P, Beule L (2021) \u0026lsquo;SRS\u0026rsquo; R Package and \u0026lsquo;q2-srs\u0026rsquo; QIIME 2 Plugin: Normalization of Microbiome Data Using Scaling with Ranked Subsampling (SRS) Applied Sciences, vol. 11\u003c/li\u003e\n \u003cli\u003eHsieh TC, Ma KH, Chao A (2016) iNEXT: an R package for rarefaction and extrapolation of species diversity (Hill numbers). Methods Ecol Evol 7(12):1451\u0026ndash;1456. https://doi.org/10.1111/2041-210X.12613\u003c/li\u003e\n \u003cli\u003eJi Y, Ashton L, Pedley SM, Edwards DP, Tang Y, Nakamura A, Kitching R, Dolman PM, Woodcock P, Edwards FA, Larsen TH, Hsu WW, Benedick S, Hamer KC, Wilcove DS, Bruce C, Wang X, Levi T, Lott M, Emerson BC, Yu DW (2013) Reliable, verifiable and efficient monitoring of biodiversity via metabarcoding. Ecol Lett 16(10):1245\u0026ndash;1257. https://doi.org/10.1111/ele.12162\u003c/li\u003e\n \u003cli\u003eKatoh K, Standley DM (2013) MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol Biol Evol 30(4):772\u0026ndash;780. http://dx.doi.org/10.1093/molbev/mst010\u003c/li\u003e\n \u003cli\u003eLeray M, Yang JY, Meyer CP, Mills SC, Agudelo N, Ranwez V, Boehm JT, Machida RJ (2013) A new versatile primer set targeting a short fragment of the mitochondrial COI region for metabarcoding metazoan diversity: application for characterizing coral reef fish gut contents. Frontiers in Zoology 10(1):34. https://doi.org/10.1186/1742-9994-10-34\u003c/li\u003e\n \u003cli\u003eLewinsohn TM, Roslin T (2008) Four ways towards tropical herbivore megadiversity. Ecol Lett 11(4):398\u0026ndash;416. https://doi.org/10.1111/j.1461-0248.2008.01155.x\u003c/li\u003e\n \u003cli\u003eLi M, Lei T, Wang G, Zhang D, Liu H, Zhang Z (2023) Monitoring insect biodiversity and comparison of sampling strategies using metabarcoding: a case study in the Yanshan Mountains, China. Ecol Evol 13(4):e10031. https://doi.org/10.1002/ece3.10031\u003c/li\u003e\n \u003cli\u003eLiu M, Clarke LJ, Baker SC, Jordan GJ, Burridge CP (2020) A practical guide to DNA metabarcoding for entomological ecologists. Ecol Entomol 45(3):373\u0026ndash;385. https://doi.org/10.1111/een.12831\u003c/li\u003e\n \u003cli\u003eMacedo-Reis LE, Quesada M, de Siqueira Neves F (2019) Forest cover drives insect guild diversity at different landscape scales in tropical dry forests. For Ecol Manage 443:36\u0026ndash;42. https://doi.org/10.1016/j.foreco.2019.04.007\u003c/li\u003e\n \u003cli\u003eMagoga G, Forni G, Brunetti M, Meral A, Spada A, De Biase A, Montagna M (2022) Curation of a reference database of COI sequences for insect identification through DNA metabarcoding: COins. Database 2022:baac055. https://doi.org/10.1093/database/baac055\u003c/li\u003e\n \u003cli\u003eMartin M (2011) Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnetjournal; Vol 17, No 1: Next Generation Sequencing Data Analysis. https://doi.org/10.14806/ej.17.1.200\u003c/li\u003e\n \u003cli\u003eMcMurdie PJ, Holmes S (2013) phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS One 8(4):e61217. https://doi.org/10.1371/journal.pone.0061217\u003c/li\u003e\n \u003cli\u003eMoreau CS, Wray BD, Czekanski-Moir JE, Rubin BER (2013) DNA preservation: a test of commonly used preservatives for insects. Invertebrate Systematics 27(1):81\u0026ndash;86. https://doi.org/10.1071/IS12067\u003c/li\u003e\n \u003cli\u003eNurul-Ain E, Rosli H, Kingston T (2017) Resource availability and roosting ecology shape reproductive phenology of rain forest insectivorous bats. Biotropica 49(3):382\u0026ndash;394. https://doi.org/10.1111/btp.12430\u003c/li\u003e\n \u003cli\u003eO\u0026apos;Rourke DR, Bokulich NA, Jusino MA, MacManes MD, Foster JT (2020) A total crapshoot? evaluating bioinformatic decisions in animal diet metabarcoding analyses. Ecol Evol 10(18):9721\u0026ndash;9739. https://doi.org/10.1002/ece3.6594\u003c/li\u003e\n \u003cli\u003eO\u0026apos;Rourke DR, Mangan MT, Mangan KE, Bokulich NA, MacManes MD, Foster JT (2021) Lord of the diptera (and moths and a spider): molecular diet analyses and foraging ecology of Indiana bats in Illinois. Frontiers in Ecology and Evolution 9. https://doi.org/10.3389/fevo.2021.623655\u003c/li\u003e\n \u003cli\u003eO\u0026rsquo;Rourke D, Rouillard NP, Parise KL, Foster JT (2022) Spatial and temporal variation in New Hampshire bat diets. Sci Rep 12(1):14334. https://doi.org/10.1038/s41598-022-17631-z\u003c/li\u003e\n \u003cli\u003ePag\u0026egrave;s H, Aboyoun P, Gentleman R, DebRoy S (2024) Biostrings: Efficient manipulation of biological strings\u003c/li\u003e\n \u003cli\u003ePatrick HJH, Chomič A, Armstrong KF (2016) Cooled propylene glycol as a pragmatic choice for preservation of DNA from remote field-collected diptera for next-generation sequence analysis. J Econ Entomol 109(3):1469\u0026ndash;1473. https://doi.org/10.1093/jee/tow047\u003c/li\u003e\n \u003cli\u003ePearson WR (2013) An introduction to sequence similarity (\u0026ldquo;homology\u0026rdquo;) searching. Current Protocols in Bioinformatics 42(1):3.1.1\u0026ndash;3.1.8. https://doi.org/10.1002/0471250953.bi0301s42\u003c/li\u003e\n \u003cli\u003ePhillips CD, Hanson J, Wilkinson JE, Koenig L, Rees E, Webala P, Kingston T (2017) Microbiome structural and functional interactions across host dietary niche space. Integr Comp Biol 57(4):743\u0026ndash;755. https://doi.org/10.1093/icb/icx011\u003c/li\u003e\n \u003cli\u003ePosit team (2023) RStudio: integrated development environment for R. Posit Software, PBC., Boston, MA\u003c/li\u003e\n \u003cli\u003eQuinn GP, Keough MJ (2002) Experimental design and data analysis for biologists. Cambridge University Press, Cambridge\u003c/li\u003e\n \u003cli\u003eR Core Team (2023) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria\u003c/li\u003e\n \u003cli\u003eRatnasingham S, Hebert PDN (2007) bold: the barcode of life data system (http://www.barcodinglife.org). Mol Ecol Notes 7(3):355\u0026ndash;364. https://doi.org/10.1111/j.1471-8286.2007.01678.x\u003c/li\u003e\n \u003cli\u003eRobeson MS, II, O\u0026rsquo;Rourke DR, Kaehler BD, Ziemski M, Dillon MR, Foster JT, Bokulich NA (2021) RESCRIPt: Reproducible sequence taxonomy reference database management. PLoS Comput Biol 17(11):e1009581. https://doi.org/10.1371/journal.pcbi.1009581\u003c/li\u003e\n \u003cli\u003eRoslin T, Somervuo P, Pentinsaari M, Hebert PDN, Agda J, Ahlroth P, Anttonen P, Aspi J, Blagoev G, Blanco S, Chan D, Clayhills T, deWaard J, deWaard S, Elliot T, Elo R, Haapala S, Helve E, Ilmonen J, Hirvonen P, Ho C, It\u0026auml;mies J, Ivanov V, Jakovlev J, Jusl\u0026eacute;n A, Jussila R, Kahanp\u0026auml;\u0026auml; J, Kaila L, Jari P, Kakko A, Kakko I, Karhu A, Karjalainen S, Kjaerandsen J, Koskinen J, Laasonen EM, Laasonen L, Laine E, Lampila P, Levesque-Beaudin V, Lu L, L\u0026auml;hteenaro M, Majuri P, Malmberg S, Manjunath R, Martikainen P, Mattila J, McKeown J, Mets\u0026auml;l\u0026auml; P, Miklasevskaja M, Miller M, Miskie R, Muinonen A, Veli M, Naik S, Nikolova N, Nupponen K, Ovaskainen O, \u0026Ouml;sterblad I, Paasivirta L, Pajunen T, Parkko P, Paukkunen J, Penttinen R, Perez K, Pohjoism\u0026auml;ki J, Prosser S, Raekunnas M, Rahulan M, Rannisto M, Ratnasingham S, Raukko P, Rinne A, Rintala T, Miranda Romo S, Salmela J, Salokannel J, Savolainen R, Schulman L, Sihvonen P, Soliman D, Sones J, Steinke C, St\u0026aring;hls G, Tabell J, Tiusanen M, V\u0026aacute;rkonyi G, Vesterinen EJ, Viitanen E, Vikberg V, Viitasaari M, Vilen J, Warne C, Wei C, Winqvist K, Zakharov E, Mutanen M (2022) A molecular-based identification resource for the arthropods of Finland. Mol Ecol Resour 22(2):803\u0026ndash;822. https://doi.org/10.1111/1755-0998.13510\u003c/li\u003e\n \u003cli\u003eRoswell M, Dushoff J, Winfree R (2021) A conceptual guide to measuring species diversity. Oikos 130(3):321\u0026ndash;338. https://doi.org/10.1111/oik.07202\u003c/li\u003e\n \u003cli\u003eS\u0026aacute;nchez-Bayo F, Wyckhuys KAG (2019) Worldwide decline of the entomofauna: a review of its drivers. Biol Conserv 232:8\u0026ndash;27. https://doi.org/10.1016/j.biocon.2019.01.020\u003c/li\u003e\n \u003cli\u003eShen W, Le S, Li Y, Hu F (2016) SeqKit: A cross-platform and ultrafast toolkit for fasta/q file manipulation. PLoS One 11(10):e0163962. https://doi.org/10.1371/journal.pone.0163962\u003c/li\u003e\n \u003cli\u003eShimoda M, Honda K-i (2013) Insect reactions to light and its applications to pest management. Applied Entomology and Zoology 48(4):413\u0026ndash;421. https://doi.org/10.1007/s13355-013-0219-x\u003c/li\u003e\n \u003cli\u003eStork NE (2018) How many species of insects and other terrestrial arthropods are there on earth? Annu Rev Entomol 63(1):31\u0026ndash;45. https://doi.org/10.1146/annurev-ento-020117-043348\u003c/li\u003e\n \u003cli\u003eTaberlet P, Coissac E, Pompanon F, Brochmann C, Willerslev E (2012) Towards next-generation biodiversity assessment using DNA metabarcoding. Mol Ecol 21(8):2045\u0026ndash;2050. https://doi.org/10.1111/j.1365-294X.2012.05470.x\u003c/li\u003e\n \u003cli\u003eWickham H (2007) Reshaping data with the reshape package. Journal of Statistical Software 21(12):1 \u0026ndash; 20. https://doi.org/10.18637/jss.v021.i12\u003c/li\u003e\n \u003cli\u003eWickham H, Averick M, Bryan J, Chang W, McGowan LDA, Fran\u0026ccedil;ois R, Grolemund G, Hayes A, Henry L, Hester J (2019) Welcome to the tidyverse. Journal of Open Source Software 4(43):1686. https://doi.org/10.21105/joss.01686\u003c/li\u003e\n \u003cli\u003eYoung RG, Mili\u0026aacute;n-Garc\u0026iacute;a Y, Yu J, Bullas-Appleton E, Hanner RH (2021) Biosurveillance for invasive insect pest species using an environmental DNA metabarcoding approach and a high salt trap collection fluid. Ecol Evol 11(4):1558\u0026ndash;1569. https://doi.org/10.1002/ece3.7113\u003c/li\u003e\n \u003cli\u003eYu DW, Ji Y, Emerson BC, Wang X, Ye C, Yang C, Ding Z (2012) Biodiversity soup: metabarcoding of arthropods for rapid biodiversity assessment and biomonitoring. Methods Ecol Evol 3(4):613\u0026ndash;623. https://doi.org/10.1111/j.2041-210X.2012.00198.x\u003c/li\u003e\n \u003cli\u003eZakaria N, Senawi J, Musa FH, Belabut D, Onn CK, Nor SM, Ahmad N (2014) Species composition of amphibians and reptiles in Krau Wildlife Reserve, Pahang, Peninsular Malaysia. Check List 10(2):335\u0026ndash;343. https://doi.org/10.15560/10.2.335\u003c/li\u003e\n \u003cli\u003eZar JH (2010) Biostatistical analysis, 5th Edition edn. Prentice Hall, Inc., Upper Saddle River, New Jersey\u003c/li\u003e\n \u003cli\u003eZenker MM, Specht A, Fonseca VG (2020) Assessing insect biodiversity with automatic light traps in Brazil: pearls and pitfalls of metabarcoding samples in preservative ethanol. Ecol Evol 10(5):2352\u0026ndash;2366. https://doi.org/10.1002/ece3.6042\u003c/li\u003e\n \u003cli\u003eZhang J, Kobert K, Flouri T, Stamatakis A (2014) PEAR: a fast and accurate Illumina paired-end read merger. Bioinformatics 30(5):614\u0026ndash;620. http://dx.doi.org/10.1093/bioinformatics/btt593\u003c/li\u003e\n \u003cli\u003eZizka VMA, Geiger MF, H\u0026ouml;rren T, Kirse A, Noll NW, Sch\u0026auml;ffler L, Scherges AM, Sorg M (2022) Repeated subsamples during DNA extraction reveal increased diversity estimates in DNA metabarcoding of malaise traps. Ecol Evol 12(11):e9502. https://doi.org/10.1002/ece3.9502\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"insect homogenate, subsample, database, sample coverage, blast","lastPublishedDoi":"10.21203/rs.3.rs-7367951/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7367951/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eMetabarcoding is widely used for molecular identification of organisms. Due to its efficiency for \u003cem\u003een masse\u003c/em\u003e characterization of samples, the technique is useful for insect biodiversity surveys. Although metabarcoding has been used for nearly two decades, there is still a need for optimized insect sample processing strategies. The goal of this study was to establish best practices for molecular characterization of bulk insect collections. We sampled insect diversity using light traps in lowland dipterocarp forest of Tengku Hassanal Wildlife Reserve in Pahang, Malaysia. Each light trap sample was homogenized and repeatedly subsampled to identify the number of subsamples required to detect total estimated insect OTU diversity in a light trap. Insect OTU diversity from 72 subsamples (12 subsamples from each of six light traps) was characterized by sequencing part of the mitochondrial Cytochrome Oxidase I (COI) gene. We found that four and eight 100 \u0026micro;L subsamples were sufficient to detect at least 90% and 95%, respectively, of insect OTU diversity collected in each light trap sample, regardless of the range of variation in preserved biomass (137.9 ̶ 445.4 g). We also built a global database from the Barcode of Life Database (BOLD) repository for our target sequencing region (313bp). We found that standardized bitscores of tblastx were significantly higher than blastn (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001); however, the percent identity distributions were not statistically different, with both around 92\u0026ndash;93%. In addition to relatively low average match identities, we also found poor taxonomic concordance between blastn and tblastx, especially at lower taxonomic levels, and suggest the advantage of metabarcoding should be leveraged by focusing on phylogenetic diversity instead of taxonomy wherever possible.\u003c/p\u003e","manuscriptTitle":"Assessment of Insect Communities with Metabarcoding","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-04 08:06:23","doi":"10.21203/rs.3.rs-7367951/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"01afc6d4-f21d-49ad-b5c9-7f324876e71c","owner":[],"postedDate":"September 4th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-02-24T03:39:53+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-04 08:06:23","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7367951","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7367951","identity":"rs-7367951","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}