Data from: Home is where the hollow is: Revealing vertebrate tree hollow user biodiversity with eDNA metabarcoding
Data files
Sep 27, 2023 version files 6.87 GB
-
filtered_ZOTU_table.csv
-
FTP348_P_Hollows.fastq.gz
-
FTP350_Hollows.fastq.gz
-
README.md
-
Sample_barcode.xlsx
-
sample_data.csv
Abstract
Tree hollows are essential for many vertebrate species, providing both nesting sites and shelter. Globally, old hollow-bearing trees are in decline resulting in many dependent species being under threat. It is therefore imperative that vital hollow-bearing trees are preserved, but it is logistically difficult to rapidly determine which hollows are being used and by which taxa. Here, we investigate the efficacy of eDNA metabarcoding as a survey tool for vertebrate hollow users. We compared the identity and richness of hollow-inhabiting vertebrate taxa using eDNA metabarcoding of both sediment from the tree hollows, and material collected using roller swabs. Samples (n = 138) were collected from hollow-bearing tuart trees (Eucalyptus gomphocephala; N = 28), within both an urban remnant and a relatively undisturbed forested area of South-West Western Australia. We detected a wide range of vertebrate taxa, including cryptic species such as the brush-tailed phascogale (Phascogale tapoatafa), while also providing ecologically informative data, such as hollow use by invasive Rainbow Lorikeet (Trichoglossus moluccanus) within the study areas. Our results showed variation in the species detected between methods, with the roller swab method detecting a greater number of species and a higher mean species richness per sample than hollow sediment did. The species detected from both methods did not perfectly overlap, highlighting the value of using multiple methods or substrates to detect a greater number of taxa. Our results suggest eDNA metabarcoding from tree hollow samples offers a sensitive and resource-efficient method of monitoring vertebrate hollow users, if enough hollows are sampled. This provides not only a broad biodiversity assessment tool but also an effective method for detecting taxa that may be elusive using other methods.
README: Data from: Home is where the hollow is: Revealing vertebrate tree hollow user biodiversity with eDNA metabarcoding
https://doi.org/10.5061/dryad.c59zw3r9r
Description of the data and file structure
There are five files included:
FTP350_Hollows.fastq.gz - Compressed sequencing file containing raw sequences from the illumina MiSeq run
FTP348_P_Hollows.fastq.gz - Compressed sequencing file containing raw sequences from the illumina MiSeq run
Sample_barcode.xlsl - The information required to demultiplex raw sequences
sample data.csv- Sample metadata for the project
filtered_ZOTU_table.csv - final ZOTU table produced after using eDNAFlow bioinformatics pipeline for sequence analysis. The taxonomy is included in this file and the details on the bioinformatics steps performed to produce this file are further described in the methods section of the publications and the dryad digital repository.
Details for Sample_barcode.xlsl
MiSeq_Run ID of the MiSeq run
MiSeq_Run_Date Date of the sequencing run
Sample.ID Sample ID
Description The type of sample taken from hollow (Roller swab/sediment)
Site The sample site (BP = Bold Park, KP = Kings Park, TNP_North = Tuart National Park North, TNP_South = Tuart National Park South)
Extraction The extraction kit used
F_Tag_ID ID of the Forward multiplex identifier (MID) tag
R_Tag_ID ID of the reverse multiplex identifier (MID) tag
F_Primer, R_Primer IDs of the forward and reverse primers
F_Tag_seq sequence of the Forward multiplex identifier (MID) tag
R_Tag_seq sequence of the Reverse multiplex identifier (MID) tag
F_Primer_Seq Forward Primer sequence
R_Primer_Seq Reverse Primer sequence
Primer Name for the assay
Details for sample data.csv
Sample Sample ID
Site The sample site (BP = Bold Park, KP = Kings Park, TNP_North = Tuart National Park North, TNP_South = Tuart National Park South)
Status Site classification, Urban reminant (Urban) and National Park
Location The location of the sample sites
tree_num Tree number sampled
hollow_num Hollow number sampled (per tree)
sample_type The type of sample taken from hollow (Roller swab/sediment)
Height Height to base of hollow sampled
Diameter Diameter of hollow sampled
Depth Depth to base of hollow sampled from hollow entrance
Methods
Study sites and Sample collection
Samples were collected between September 2020 and January 2021 from a total of 28 tuart trees, including 14 trees from two sites in the Tuart Forest National Park, Ludlow (hollows = 37, total samples = 58, roller = 37, sediment = 21), and 14 trees from two urban remnants within the City of Perth, Kings Park and Bold Park (hollows = 56, total samples = 80, roller = 56, sediment = 24; Figure 1).
At each site, seven hollow-bearing trees were chosen due to their accessibility and likelihood of containing at least one potential hollow based on visual surveys, resulting in a total of 93 hollows sampled. Within the urban remnants of Kings Park (September/November 2020) and Bold Park (December 2020), an elevated work platform was used to access tree hollows. Single rope access techniques were used to access trees in the Tuart Forest National Park (January 2021). For all hollows (n = 93), a roller swab sample, adapted from Valentin et al. (2020), was taken using a 40 mm polyester roller dampened with distilled water and stored in a zip lock bag. Prior to sampling, each hollow was inspected and if an animal was present, only the opening and immediate inside walls were swabbed. For all other hollows, the opening, inside walls, and across the hollow floor were all sampled. Where hollow sediment was present and accessible (n = 45), a 50 ml sterile container was used to collect surface sediment randomly from multiple locations across each hollow floor, creating one bulk sediment sample per hollow. Care was taken not to sample sediment disturbed by the roller sampling. Samples were then stored on ice, frozen within 48hrs of sampling, and stored at -18 °C prior to extraction. Visual vertebrate activity was also recorded (e.g., animal present, feathers, fur, eggs). Finally, the location, height to hollow opening, hollow opening diameter, depth of hollow, and tree width at breast height were recorded.
To prevent contamination of samples, sterile gloves were worn during sampling and single-use disposable, pre-sterilized collection containers were used. Single-use polyester rollers were decontaminated prior to use by submerging in a 10 % bleach solution (made using White King Premium Bleach, ~ 5 % sodium hypochlorite) for a minimum of 10 minutes before being thoroughly rinsed with purified distilled water and UV-sterilized for a further 10 minutes. All multi-use field sampling equipment such as roller frames were decontaminated onsite prior to use by submerging in 100 % bleach (White King Premium Bleach, ~ 5 % sodium hypochlorite) for a minimum of 10 minutes before being thoroughly rinsed with distilled water. Multi-use equipment was also swabbed prior to use before being preserved and processed in the same way as field samples as a negative control allowing for the detection of possible contamination of samples.
Sample processing and DNA extraction
Sample processing and extraction were undertaken in dedicated laboratories at the Trace & Environmental DNA (TrEnD) Laboratory, Curtin University, Perth, Western Australia. Each individual roller swab was placed in a plastic container decontaminated with 10 % bleach solution as outlined above, covered with purified distilled water (~ 500 mL) for 10 minutes and manually agitated twice for 30 seconds at 5-minute intervals. Water samples were then individually filtered across Pall 0.45μm GN-6 Metricel® mixed cellulose ester membranes using a peristaltic Pall Sentino® Microbiology pump (Pall Corporation, Port Washington, USA) and frozen at -18°C prior to extraction. DNA was extracted from half of the filter membrane using a DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany), with the following modifications to the standard protocol: 60 μl of Proteinase K, 540 μl of ATL lysis buffer and a 17-hour digestion at 56°C. Sediment samples were manually shaken for 30 seconds to homogenise and DNA was extracted from 0.25 g of sample using a DNeasy PowerLyzer PowerSoil Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions. Blank extraction controls (n = 6) containing reagents only were processed concurrently with samples to detect any cross-contamination and eDNA extracts were then stored at -18°C.
Assessment of DNA extracts
In silico studies were conducted using DNA sequences from the 12S and 16S rRNA marker regions (Othman et al., 2021) to identify the assay able to amplify the widest range of vertebrate taxa in the study region. The primer set 12S-V5 F2/R2 (Riaz et al., 2011), previously used to successfully amplify vertebrate taxa within the study region (Ryan, Bateman, van der Heyde, Fernandes, & Nevill, 2020; Van Der Heyde et al., 2021; Van Der Heyde et al., 2020), was broad enough to amplify avian, mammalian and reptilian DNA, and short enough to pick up degraded DNA. As such, we amplified DNA with the general vertebrate primer 12S-V5 (Riaz et al., 2011), designed to amplify a 98bp fragment which includes the V5 variable region of the 12S rRNA gene with the primers F2 (5’-TAGAACAGGCTCCTCTAG-3’) and R2 (5’-TTAGATACCCCACTATGC-3’).
Quantitative PCR (qPCR) was used to assess the quality and quantity of DNA in the extracts, as well as potential inhibition to determine the optimal dilution level of DNA input (Murray, Coghlan, & Bunce, 2015). All qPCR reactions were prepared in dedicated clean rooms at the TrEnD Laboratories and performed on a StepOnePlusTM Real-Time PCR system (Applied Biosystems, Massachusetts, USA). Reaction volumes totalled 25μl containing 1× PCR Gold buffer (Applied Biosystems), 2.5 mM MgCl2 (Applied Biosystems), 0.4 mg/ml BSA (Fisher Biotec, Australia), 0.25 mM of each dNTPs (Astral Scientific Australia), 0.4 μM forward and reverse primer, 1 U AmpliTaq Gold (Applied Biosystems), 0.6 µL SYBR Green (Life Technologies, USA) and 2 μL of template eDNA (3 PCR reactions per extract at Neat, 1:10, and 1:100 dilutions). The cycling conditions were initial denaturation for 5 minutes at 95 °C, followed by 50 cycles of 30 seconds at 95 °C, then 30 seconds at the annealing temperature of 60 °C, and 45 seconds at 72 °C, ending with 10 minutes of elongation at 72 °C. For each qPCR run, a non-template control (reagents only) and positive control (Gallus gallus, 100ng) were used to detect contamination and to ensure primers attached to DNA strands.
DNA amplification
Samples containing a sufficient quantity of amplifiable DNA (n = 76; Table S1) were assigned a unique 6-8 bp multiplex identifier tag (MID tag), 12S-V5 specific primers and Illumina sequencing adaptors. The unique combinations of MID tag (fusion) primers were then used in qPCR which was prepared in a physically separate ultra-clean laboratory with 4 μL of template eDNA for reaction volumes totalling 27μl, which contained the same reagents, controls and cycling conditions described above. A single-step fusion protocol was used with no reuse of index combinations. MID-tag amplicons were generated in duplicate to reduce the effects of PCR stochasticity (Murray et al., 2015) and combined before being pooled with other MID-tag amplicons (i.e. minipool) according to qPCR ΔRn values. Minipool concentrations (ng/µl) were quantified using a QiAxcel Advanced System (Qiagen, Hilden, Germany) before being combined in approximate equimolar ratios based on this quantification to create a DNA sequencing library which also contained one unrelated project. Amplicons in this library were size selected to 150 – 300 bp using a PippinPrep 2 % ethidium bromide cassette (Sage Science, Beverly, USA), purified using a QIAquick PCR Purification Kit (Qiagen, Hilden, Germany), and re-quantified using a Qubit Fluorometer (Invitrogen, Carlsbad, USA). Sequencing was then undertaken on an Illumina MiSeq platform (Illumina, San Diego, USA) using a single lane flow cell as per Illumina protocols for single-end sequencing (MiSeq® v2 Reagent Kit 300 Cycles PE), with a final library molarity of 6pM containing 7 % PhiX.
Sequence analysis (filtering and taxonomic assignment)
Conservative filtering steps were employed whilst also attempting to retain as much of the “true” eDNA diversity as possible. All sequence filtering, ZOTU formation and taxonomic assignment were conducted utilizing the eDNAFlow bioinformatics pipeline (Mousavi‐Derazmahalleh et al., 2021) via Zeus, an SGI cluster, based at the Pawsey Supercomputing Centre in Kensington, Western Australia. Sequence reads were quality checked using FASTQC (Andrews, 2010), and quality filtered (Phred quality score <20), including trimming sequences with Ns as implemented in eDNAFlow source code by --trimns & --trimqualities employing AdapterRemoval v2 (Schubert, Lindgreen, & Orlando, 2016). Demultiplexing with obitools (ngsfilter; Boyer et al., 2016) and sequences under 60 bp removed (obigrep). Dereplication and the creation of zero-radius operational taxonomic units (ZOTUs; min sequence abundance = 4) and ZOTU tables were then undertaken using the USEARCH unoise3 algorithm (Edgar, 2016). The ZOTUs were queried using BLASTN (Altschul, Gish, Miller, Myers, & Lipman, 1990) against NCBI’s GenBank nucleotide database using the following parameters: % identity ≥ 95, evalue ≤ 1e−3, % query cover 97, max target sequences = 10. Erroneous ZOTUs derived by high-throughput sequencing were identified and removed using the post-clustering curation method LULU with the minimum threshold of sequence similarity at 95 % (Frøslev et al., 2017). Finally, curated ZOTUs were taxonomically assigned by the lowest common ancestor (LCA) approach using eDNAFlow custom python script (Mousavi‐Derazmahalleh et al., 2021), with taxonomic assignment collapsed to the LCA if the percent identity of two hits with 97 % query cover and 95 % identity, differed by less than 1 % (Table S2).
Further sample filtering was performed in R v.4.0.2 (R Core Team, 2018) using the “phyloseq” package (McMurdie & Holmes, 2013). To mitigate against cross-contamination and false positives, analysis of contamination present in negative controls, and rarefaction curve (Figure S1), resulted in taxa with sequence counts fewer than 3 and samples with low sequencing depth (<50 reads), being removed. Following this, sequences present in negative/extraction controls, taxa not known to inhabit or interact with hollows or be the prey of those that may, common PCR contaminants (e.g., human, and ungulate sequences) and taxonomic assignments higher than family level were removed from the dataset.
Taxonomic assignments were then corrected based on existing species distribution data for each site. Species Red-Collared Lorikeet (Trichoglossus rubritorquis), short-eared possum (Trichosurus caninus) and New Zealand Fantail (Rhipidura fuliginosa), which are not known to occur within SWWA, were reassigned to local taxa Rainbow Lorikeet (Trichoglossus moluccanus), brushtail possum (Trichosurus vulpecula) and Grey Fantail (Rhipidura albiscapa), respectively. The genus Tadorna and family Pseudocheiridae were reassigned to Australian Shelduck (Tadorna tadornoides) and western ringtail possum (Pseudocheirus occidentalis), respectively, due to these taxa containing only a single species in SWWA. The family Strigidae was reassigned to genus Ninox again due to known species distributions; however, this could not be assigned to a lower level as two Ninox species are known to occur within the SWWA. The Malayan field rat (Rattus tiomanicus) and Himalayan field rat (Rattus nitidus) were downgraded to genus level, due to species distributions outside of the study area. ZOTUs with multiple possible assignments were left at the closest taxonomic level and labelled “sp”. Of these Pstittacidae sp. was reassigned to Psittaculidae sp. in line with species distribution data.