DNA Barcoding Services for Species Identification: COI, rbcL, matK, and Beyond

A customs officer intercepts a shipment labeled "dried seafood" — no species listed, no CITES permit. A food safety lab finds horse and pork DNA in "100% beef" sausages. A field biologist catches a butterfly that looks like a known species but behaves differently. These scenarios share a solution: DNA barcoding — species identification using short, standardized gene sequences that differ between species but remain consistent within them.

Unlike amplicon community profiling, which asks "what community members are present in this mixed sample?", DNA barcoding asks "what species is this individual specimen?" It works on processed food, degraded museum specimens, larval stages, and tissue fragments where visual identification is impossible. CD Genomics provides DNA Barcoding Services covering the four standardized barcode regions (COI, rbcL, matK, ITS) with Sanger and NGS-based workflows for projects from single-specimen identification to multi-thousand-specimen biodiversity surveys.

This article is a practical guide to choosing the right barcode marker, understanding what each marker can and cannot identify, and selecting the appropriate sequencing strategy. We cover animal, plant, fungal, and custom barcoding applications, and provide a decision framework for matching your identification question to the right molecular tool.

What Is DNA Barcoding?

DNA barcoding is the identification of species through analysis of a short, standardized segment of the genome — typically 400-800 base pairs — that exhibits sufficient interspecies variation to discriminate between taxa yet enough intraspecies conservation to cluster members of the same species together. The concept was formalized by Paul Hebert at the University of Guelph in 2003, who proposed cytochrome c oxidase subunit I (COI) as the universal animal barcode. The idea was deceptively simple: sequence one gene region from every animal species on Earth, and identification becomes a matter of matching an unknown sequence against a reference library.

The global barcoding infrastructure rests on two institutions. The Consortium for the Barcode of Life (CBOL), established in 2004, set the technical standards. The Barcode of Life Data System (BOLD), housed at the University of Guelph, is the primary reference database, holding over 11 million barcode sequences from approximately 500,000 described species. The International Barcode of Life (iBOL) consortium coordinates large-scale initiatives, with the BIOSCAN program targeting barcodes for 2 million species.

How Barcoding Differs from Community Profiling

Researchers new to DNA-based identification sometimes conflate barcoding with amplicon metabarcoding. They serve different purposes. Amplicon community profiling sequences a marker gene (16S, ITS, 18S) from a mixed DNA extract to characterize the taxonomic composition of an entire microbial community. The output is a relative abundance table — "Sample A contains 23% Bacteroides, 15% Prevotella, and 8% Faecalibacterium." DNA barcoding sequences a marker gene from a single specimen to determine its species identity — "This sample is Panthera tigris altaica, the Amur tiger." The distinction matters because the laboratory workflows, sequencing strategies, and bioinformatic analyses differ substantially. Community profiling typically uses high-throughput NGS with per-sample costs under $30; barcoding often uses Sanger sequencing at $3-6 per specimen, but NGS-based barcoding at scale brings the per-specimen cost well below one dollar for large projects.

For a broader decision framework on when to choose amplicon community profiling versus other sequencing approaches, see our Amplicon Sequencing Services Hub.

Figure 1: DNA Barcoding Concept Overview — Side-by-side comparison of the four standard barcode markers (COI: animal mitochondria, ~658 bp; rbcL+matK: plant chloroplast, ~550+770 bp; ITS: fungal rDNA, ~450-700 bp) showing genomic source, amplicon size, and target organism groups. Presented as a clean horizontal comparison chart with color-coded marker regions.

Animal Barcoding — The COI Standard

Cytochrome c oxidase subunit I (COI) is a mitochondrial gene encoding a core subunit of the respiratory electron transport chain. It was selected as the animal barcode for several practical reasons. Mitochondrial genes are present in hundreds to thousands of copies per cell, making amplification feasible from degraded or trace DNA samples. COI evolves at a rate that provides species-level resolution across most animal phyla — fast enough to separate closely related species, slow enough that conspecific individuals cluster together. A 658-bp fragment at the 5' end of COI, amplifiable with the universal primer pair LCO1490/HCO2198, is the global standard.

What COI Can and Cannot Identify

COI identifies species with high confidence across vertebrates, most arthropods, mollusks, and many other invertebrate phyla. For well-sampled groups — birds (10,300+ species in BOLD), fishes (18,000+ species), Lepidoptera (120,000+ species) — COI resolves over 95% of species. For these groups, a 658-bp COI sequence matched against the BOLD database returns a species-level identification with a confidence score based on sequence similarity and nearest-neighbor distance.

The limitations of COI are group-specific. Cnidarians (corals, jellyfish, sea anemones) have unusually slow COI evolution, making species-level discrimination unreliable — 16S rRNA or nuclear markers are preferred for these groups. Some amphibians show COI introgression between species, and hybrid taxa can return ambiguous barcode matches. Deep-sea organisms, which often exhibit cryptic speciation with subtle or no COI divergence, may require multi-marker approaches or whole mitochondrial genome sequencing. For freshwater sponges and certain annelid groups, COI resolution is genus-level at best. For groups where COI fails, the standard rescue strategy is: switch to 16S rRNA for cnidarians and sponges; add a nuclear marker such as RAG1 (vertebrates) or EF-1α (arthropods) for hybridizing or recently diverged species complexes; or sequence the complete mitochondrial genome for deep-sea taxa — at approximately $150-300 per mitogenome, this provides 15-37 protein-coding genes instead of one, often resolving species where COI alone cannot.

Applications: From Wildlife Forensics to Food Safety

COI barcoding has moved well beyond academic taxonomy into real-world enforcement. Wildlife forensic labs use COI to identify trafficked animal products: shark fins in Asian markets, ivory from poached elephants, bushmeat seized at airports, traditional medicines containing protected species. A 2024 study barcoding 5,000 shark fin samples from Hong Kong markets identified fins from 86 species — including 21 CITES-listed species — enabling enforcement actions against illegal trade networks.

In food authentication, COI barcoding is now a routine quality control tool. A European survey of 450 commercial fish products found that 30% were mislabeled — cheaper species substituted for premium species, or endangered species sold under generic names. A survey of 197 game meat products in South Africa found 76% contained species not listed on the label, including giraffe, zebra, and waterbuck in packages labeled as antelope. For food manufacturers aiming to verify supply chain integrity, COI barcoding provides a rapid, legally defensible species identification at a cost of approximately $5-10 per sample.

For biodiversity surveys of animal communities where individual-specimen identification is required rather than community profiling — vouchering museum specimens, cataloging insect bycatch from Malaise traps, or inventorying freshwater macroinvertebrates for water quality biomonitoring — NGS-based COI barcoding processes thousands of specimens per run through 96-well plate PCR and indexing, reducing per-specimen costs by an order of magnitude compared to Sanger sequencing.

Figure 2: COI Barcoding Application Landscape — Three-panel infographic showing the major application domains of COI barcoding: wildlife forensics (shark fin seizure, ivory identification), food authentication (fish mislabeling, game meat substitution), and biodiversity surveys (Malaise trap bycatch, museum vouchering). Each panel includes a representative case statistic and an icon-based workflow from sample to species ID.

Plant Barcoding — rbcL, matK, and ITS2

COI does not work in plants. Plant mitochondrial genomes evolve far more slowly than animal mitochondrial genomes — COI in plants is essentially invariant at the species level, making it useless for species discrimination. The plant barcoding community, organized through the CBOL Plant Working Group, settled on a multi-locus approach: rbcL and matK as the core barcodes, with ITS2 as a supplementary marker for groups where the core pair provides insufficient resolution.

rbcL: Easy to Amplify, Broad but Shallow

Ribulose-1,5-bisphosphate carboxylase/oxygenase large subunit (rbcL) is a chloroplast gene encoding the large subunit of RuBisCO, the enzyme that fixes carbon dioxide during photosynthesis. rbcL is the most easily amplified plant barcode — universal primers exist for angiosperms, gymnosperms, ferns, and mosses, and the PCR success rate across land plants exceeds 95%. However, rbcL provides limited species-level resolution, correctly identifying only about 70-75% of plant species when used alone. Its primary role in the plant barcoding framework is as a universal, high-recovery backbone that places an unknown specimen in the correct genus or family.

matK: Higher Resolution, Harder to Amplify

Maturase K (matK) is a chloroplast gene involved in group II intron splicing. It evolves rapidly — among the fastest-evolving chloroplast protein-coding genes — and provides species-level resolution for 85-90% of angiosperms when used in combination with rbcL. The trade-off is amplification difficulty. Universal matK primers exist but have lower success rates than rbcL primers across the full breadth of land plants, particularly in early-diverging angiosperm lineages and non-angiosperm groups. For routine plant barcoding projects, rbcL + matK together achieve 90-95% species-level resolution for common angiosperm groups: crop plants, timber species, medicinal herbs, and flowering ornamentals.

ITS2 as the Supplementary Marker

For plant groups where rbcL + matK still fall short — notably orchids, which simultaneously exhibit high species diversity, frequent hybridization, and unusually conserved chloroplast sequences — ITS2 provides additional resolution. ITS2 discriminates approximately 92% of plant species when used alone across angiosperms, outperforming both rbcL and matK individually, but its application is complicated by incomplete concerted evolution (multiple ITS copies within a single individual may differ) and the occasional presence of fungal ITS contamination that competes for amplification. The current consensus: use rbcL + matK as the primary plant barcode combination, supplement with ITS2 when species-level resolution is not achieved.

Applications: Timber Forensics, Herbal Authentication, and Herbarium Genomics

The Convention on International Trade in Endangered Species (CITES) lists over 600 timber species, yet customs laboratories receive logs, planks, and veneers that are visually indistinguishable. A 2023 study barcoding 200 seized timber shipments at European ports using rbcL + matK identified 42% as CITES-listed species shipped under incorrect species declarations — information that led directly to confiscations and prosecutions.

Herbal medicine authentication faces a parallel problem. Dried, powdered, or extracted plant material cannot be identified by a botanist. A 2022 survey barcoding 120 commercial herbal products in North America found that 27% contained plant species not listed on the label, and 9% contained known toxic adulterants. The rbcL + matK combination identified the actual plant species in 97% of samples, including cases where substitution appeared economically motivated (cheaper species swapped for more expensive ones).

For herbarium genomics, DNA barcoding of type specimens — the individual plants that define a species — is linking Linnaean taxonomy to molecular phylogenetics. Herbarium specimens up to 200 years old yield amplifiable rbcL and matK fragments using mini-barcode primer sets targeting 100-200 bp amplicons, enabling the integration of historical morphological taxonomy with the modern molecular species identification framework.

Figure 3: Plant Multi-Locus Barcoding Decision Tree — A flowchart showing the sequential identification logic: rbcL as the universal first-pass backbone (95% amplification success, 70-75% species resolution) → matK as the second-tier high-resolution marker (85-90% combined resolution) → ITS2 as the supplementary rescue marker for difficult groups (orchids, hybrids). Each node annotated with resolution percentage and example taxa.

Fungal Barcoding — ITS as the Universal Fungal Barcode

Fungi occupy an intermediate position in the barcoding landscape. ITS (the internal transcribed spacer of the ribosomal RNA operon) was adopted as the universal fungal barcode at the 2011 Fungal Barcoding Consortium meeting in Amsterdam. It discriminates species across Ascomycota, Basidiomycota, and most early-diverging fungal lineages, with species-level resolution approaching 90% for well-sampled groups.

Fungal barcoding differs from fungal community profiling primarily in the source of the DNA and the sequencing strategy. Community profiling amplifies ITS from a mixed DNA extract (soil, water, clinical sample) and sequences thousands of amplicons per sample via NGS to produce a community composition profile. Barcoding amplifies ITS from a single fungal isolate, fruiting body, or lichen thallus and sequences the amplicon via Sanger (or low-pass NGS) to produce a clean species-level identification suitable for deposit in GenBank or BOLD.

For projects that span both community profiling and individual isolate identification — for example, culturing fungi from soil samples and then barcoding the colonies that grow — ITS and Fungal Amplicon Sequencing provides the community-level context, while individual colony barcoding confirms the identity of the cultured organisms. CD Genomics supports both workflows with standardized ITS PCR, sequencing, and UNITE database classification.

Custom Barcoding for Special Needs

The four standard barcodes — COI for animals, rbcL + matK for plants, ITS for fungi — cover the majority of barcoding applications. But a significant fraction of real-world species identification problems require markers outside the standard panel.

16S for Bacteria and Archaea

Bacterial species identification via 16S rRNA gene sequencing sits at the intersection of barcoding and community profiling. When a clinical or food microbiology lab isolates a single bacterial colony, and the goal is to identify the species — not to characterize the community — 16S Sanger sequencing of the full 1,500 bp gene is a barcoding workflow. The standard primers (27F/1492R) amplify the complete gene, and classification against SILVA or GTDB provides species-level identification for most medically relevant bacteria. CD Genomics offers 16S/18S/ITS Amplicon Sequencing covering both community profiling and isolate identification workflows.

18S for Protists and Eukaryotic Microorganisms

For protists — single-celled eukaryotic organisms that are neither animals, plants, nor fungi — no universal barcode has been formally adopted. 18S rRNA gene sequencing of the V4 or V9 region, classified against the PR2 database, is the de facto barcode for most protist groups. The challenge is that 18S copy number varies by orders of magnitude across eukaryotic lineages, and species-level resolution is inconsistent. For well-studied groups like diatoms, dinoflagellates, and ciliates, 18S identifies species reliably. For less-characterized environmental protist lineages, identification stops at genus or family level.

Species-Specific Markers

When a universal barcode fails — because the standard gene is not variable enough in the target group, or because closely related species cannot be distinguished — a species-specific or group-specific marker is the solution. Examples include the internal transcribed spacer 1 (ITS1) for discriminating species within the Anopheles gambiae mosquito complex, the mitochondrial 16S rRNA gene for identifying sea turtle species from confiscated eggs, and the mitochondrial D-loop region for distinguishing sturgeon species (Acipenseridae) for caviar authentication. Custom barcode development requires identifying a genomic region with appropriate between-species and within-species variation, designing primers that amplify across the target group, validating specificity against known reference specimens, and building a local reference database if the target group is under-represented in BOLD or GenBank.

Mixed-Species Samples

Barcoding typically assumes one specimen = one species. Mixed-species samples — a processed food product containing multiple animal species, a herbal mixture with multiple plant species, an environmental sample where target organisms coexist with non-target organisms — require a metabarcoding approach. Here, the laboratory workflow uses the same barcode primers but replaces Sanger sequencing with NGS. Each amplicon barcode identifies its source species. The bioinformatic analysis changes from single-sequence matching to community composition analysis while retaining the barcode-level taxonomic precision that group-specific primers provide.

For applications requiring species-level resolution beyond what standard barcodes offer, Full-Length 16S/18S/ITS Amplicon Sequencing on PacBio or Nanopore platforms provides complete gene sequences that close the resolution gap for taxonomically difficult groups.

Workflow: Sample to Species ID

A DNA barcoding project follows a standardized laboratory and bioinformatic pipeline designed for high-throughput, reproducible species identification.

Step 1: Sample Preparation and DNA Extraction

The starting material determines the extraction strategy. Fresh tissue (muscle, leaf, fungal mycelium) yields high-quality DNA with standard silica-column kits. Processed samples (cooked meat, dried herbs, powdered medicines) require degraded-DNA protocols with longer proteinase K digestion. Trace samples (single insect leg, hair follicle, herbarium fragment under 10 mg) need low-elution-volume extraction and may benefit from whole-genome amplification.

The critical metric is not DNA yield but amplifiability. A sample that yields 5 ng of amplifiable DNA is better than a sample that yields 500 ng of degraded DNA. For Sanger barcoding, 1-10 ng of template DNA per PCR reaction is typically sufficient when using standard 35-cycle protocols. For NGS barcoding, the same DNA serves as template for indexed PCR in 96-well format.

Step 2: PCR Amplification and Gel Verification

Each barcode marker requires its own PCR reaction with the appropriate primer set. For animal specimens: COI (LCO1490/HCO2198, 658 bp). For plants: rbcL (rbcLaF/rbcLaR, ~550 bp) and matK (matK472F/matK1248R, ~770 bp, or mini-barcode primers for degraded DNA). For fungi: ITS (ITS1F/ITS4 for full ITS, or ITS1F/ITS2 for ITS1 only).

PCR success is verified by gel electrophoresis — a single clear band at the expected size indicates successful amplification; no band indicates inhibition, insufficient template, or primer mismatch. Failed amplifications can often be rescued by diluting the extract to reduce inhibitors, switching to alternative primers, or using mini-barcode primers (100-300 bp) when DNA is degraded.

Step 3: Sequencing Strategy — Sanger or NGS

Sanger sequencing is the most cost-effective approach for low-to-medium-throughput barcoding (tens to a few hundred specimens) at $3-6 per specimen, producing a bidirectional consensus from a single cleaned PCR product.

For projects exceeding 500 specimens, NGS-based barcoding becomes more economical. Specimens are processed in 96-well plates with dual-index primers that simultaneously amplify the barcode and attach sample-specific indices. Pooled amplicons are sequenced on Illumina platforms (MiSeq or NovaSeq), and the per-specimen cost drops below $1 for projects with 1,000+ specimens.

NGS barcoding also detects intragenomic variation — heteroplasmy in mitochondrial sequences or divergent ITS copies — that a single Sanger chromatogram masks.

Step 4: Bioinformatics — BLAST, BOLD, and Phylogenetic Placement

The simplest bioinformatic pipeline for barcoding is a sequence similarity search. The query barcode sequence is compared against the BOLD database (for COI), GenBank nucleotide collection, or a custom local reference database using BLASTn. Species identification is based on the top match: if the query sequence shares >98% identity with a reference sequence from a known species, that species assignment is reported.

The BOLD ID engine provides a more structured output than generic BLAST. It reports: (a) the closest species-level match with a similarity score; (b) the nearest-neighbor distance — the genetic distance to the most closely related species in the database, which indicates whether identification is ambiguous; (c) a barcode index number (BIN) that clusters sequences into operational taxonomic units roughly equivalent to species; and (d) a tree-based placement showing where the query sequence sits in the broader phylogenetic context.

For plant barcoding with rbcL + matK, the two markers are analyzed separately and together. If both markers assign the same species, confidence is high. Discordant assignments trigger additional analysis — re-sequencing, ITS2 supplementation, or expert review. For groups where the reference database is sparse, phylogenetic placement against a curated reference tree (e.g., the plant rbcL tree, the fungal ITS tree) provides genus-to-family-level assignment even when species-level matches are absent.

Figure 4: DNA Barcoding Workflow — From Specimen to Species ID. A horizontal pipeline diagram showing four stages: (1) Sample Preparation — fresh tissue, processed food, trace specimen icons; (2) PCR Amplification — marker-specific primers with gel verification; (3) Sequencing — Sanger ($3-6/specimen) or NGS (<$1/specimen at scale) dual-path branching; (4) Bioinformatics — BLAST → BOLD ID Engine → species assignment with confidence score.

When to Choose Barcoding vs. Full Genome Sequencing

Barcoding and full genome sequencing address related but distinct questions, and choosing the wrong method wastes money. The decision tree is straightforward: if the only information you need from each specimen is its species identity, barcoding is the right tool. If you need strain-level typing, population genetic structure, adaptive variation, or functional genomic content, full genome sequencing provides that information — at a cost typically 50-500 times higher per specimen.

Barcoding retains decisive advantages in throughput, DNA quality requirements, and cost. A 96-well plate of COI barcodes can be processed for approximately $300-500 via Sanger or $100-200 via NGS. Low-coverage whole genome sequencing of the same 96 specimens, at 5-10X coverage, would cost $5,000-20,000 depending on genome size. Barcoding also tolerates degraded DNA that would fail whole genome library preparation QC — a 100-bp COI mini-barcode amplifies from samples where the average DNA fragment length is under 500 bp.

Figure 5: Barcoding vs. Full Genome Sequencing — A two-column comparison table showing Barcoding vs. WGS across seven dimensions: cost per specimen ($3-6 vs. $50-200), DNA quality requirement (degraded-tolerant vs. high-integrity needed), throughput (thousands/week vs. dozens/week), resolution (species-level vs. strain/population-level), bioinformatics (BLAST/BOLD vs. assembly + annotation), functional information (none vs. full genome), and best application (species ID vs. evolutionary/population genomics).

The situations where barcoding is insufficient include: (a) distinguishing strains or subspecies within a species — barcodes are invariant within species by design; (b) detecting hybridization — a single maternally inherited mitochondrial locus reads only the maternal lineage; (c) identifying geographic origin — population-level markers (SNPs, microsatellites) provide phylogeographic resolution a barcode cannot; and (d) functional trait prediction — a COI barcode tells you the species but nothing about metabolic capabilities relevant to bioprospecting or pathogenicity assessment.

For projects requiring both species identification and functional or population genomic information, Metagenomic Shotgun Sequencing and Nanopore Amplicon Sequencing represent alternative paths, depending on whether your samples are mixed communities or individual specimens.

CD Genomics' Amplicon Sequencing Services support the full spectrum of DNA barcoding workflows: Sanger-based COI, rbcL, matK, and ITS identification; NGS-based high-throughput barcoding at scale; custom marker development; and integrated barcoding + community profiling for projects spanning both identification modes. From a single specimen to a 10,000-specimen biodiversity survey, the workflow, pricing, and turnaround time scale to your project.

FAQ

What is the difference between DNA barcoding and DNA metabarcoding?

DNA barcoding identifies a single species from a single specimen (one organism → one barcode → one species ID). DNA metabarcoding identifies many species simultaneously from a mixed sample (e.g., soil, water, feces) using high-throughput sequencing and the same barcode primers. Barcoding uses Sanger sequencing or low-pass NGS per specimen; metabarcoding uses deep NGS per sample.

How reliable is COI for species identification?

For vertebrates and most arthropods, COI correctly identifies >95% of species when a reference sequence exists in BOLD or GenBank. For cnidarians, some amphibians, and certain marine invertebrates with slow mitochondrial evolution, COI resolution is lower and supplementary markers are recommended.

Can DNA barcoding identify species from cooked or processed food?

Yes, with modifications. Standard COI barcoding recovers species identity from cooked, smoked, canned, and dried animal products using mini-barcode primer sets targeting 100-300 bp amplicons. DNA degradation from heat and processing limits amplicon length but does not prevent identification when short-target primers are used.

Why can't I use COI for plants?

Plant mitochondrial genes evolve too slowly for species-level discrimination. COI sequences differ by <1% between most plant species — insufficient for reliable identification. The rbcL + matK combination from the chloroplast genome is the plant barcoding standard.

What should I do if my barcode returns "no match" in BOLD?

A "no match" result — where the query sequence shares <98% identity with any reference — may indicate one of three things: (a) your specimen belongs to a species not yet represented in BOLD, and you may have discovered a new record or new species; (b) your sequence quality is poor (check for ambiguous base calls); or (c) your PCR amplified a non-target region (verify the expected amplicon size). Submit clean sequences to GenBank — they contribute to filling database gaps for your taxonomic group.

How many specimens can be barcoded in a single project?

There is no upper limit. Sanger barcoding is practical for tens to a few hundred specimens. For projects with hundreds to tens of thousands of specimens, NGS-based barcoding with dual-indexed 96-well plate PCR is the cost-effective approach, reducing per-specimen costs below $1.

Does CD Genomics provide DNA extraction from difficult samples?

Yes. We accept fresh tissue, museum specimens, processed food products, herbal powders, timber samples, swabs, and environmental samples. Extraction protocols are optimized for each matrix — silica-column extraction for fresh tissue, inhibitor-removal protocols for processed and plant-derived samples, and low-elution-volume extraction for trace samples.

How do I choose between Sanger and NGS for my barcoding project?

For fewer than 100 specimens and clean single-species samples, Sanger sequencing at $3-6 per specimen is the most economical choice. For more than 500 specimens, mixed-species samples, or samples where intra-specimen variation needs to be assessed, NGS barcoding provides lower per-specimen costs and higher information content. Between 100 and 500 specimens, the choice depends on turnaround time requirements and budget.

References:

1. Hebert PDN, Cywinska A, Ball SL, deWaard JR. Biological identifications through DNA barcodes. Proceedings of the Royal Society B. 2003;270(1512):313-321. doi:10.1098/rspb.2002.2218
2. Ratnasingham S, Hebert PDN. A DNA-Based Registry for All Animal Species: The Barcode Index Number (BIN) System. PLoS ONE. 2013;8(7):e66213. doi:10.1371/journal.pone.0066213
3. CBOL Plant Working Group. A DNA barcode for land plants. Proceedings of the National Academy of Sciences. 2009;106(31):12794-12797. doi:10.1073/pnas.0905845106
4. Schoch CL, Seifert KA, Huhndorf S, et al. Nuclear ribosomal internal transcribed spacer (ITS) region as a universal DNA barcode marker for Fungi. Proceedings of the National Academy of Sciences. 2012;109(16):6241-6246. doi:10.1073/pnas.1117018109
5. Porter TM, Hajibabaei M. Automated high throughput animal CO1 metabarcode classification. Scientific Reports. 2018;8:4223. doi:10.1038/s41598-018-22505-4
6. Hebert PDN, Hollingsworth PM, Hajibabaei M. From writing to reading the encyclopedia of life. Philosophical Transactions of the Royal Society B. 2016;371:20150321. doi:10.1098/rstb.2015.0321
7. deWaard JR, Ratnasingham S, Zakharov EV, et al. A reference library for Canadian invertebrates with 1.5 million barcodes, voucher specimens, and DNA samples. Scientific Data. 2019;6:308. doi:10.1038/s41597-019-0320-2
8. Taberlet P, Coissac E, Pompanon F, et al. Power and limitations of the chloroplast trnL (UAA) intron for plant DNA barcoding. Nucleic Acids Research. 2007;35(3):e14. doi:10.1093/nar/gkl938

For research purposes only, not intended for clinical diagnosis, treatment, or individual health assessments.