DNA Barcoding, STR Profiling, and Individual Identification

Paul Hebert and colleagues at the University of Guelph published the DNA barcoding concept in 2003, proposing the 648 base-pair 5' region of cytochrome c oxidase subunit I as a universal animal identifier. The choice was not arbitrary. COI sits in mitochondrial DNA, which is present in thousands of copies per cell: useful when tissue quantity is tiny or degraded. The locus evolves fast enough that different species usually carry sequences that differ by more than 2-3%, yet slow enough that members of the same species cluster tightly. That gap between intraspecific and interspecific variation is what makes species discrimination possible.

For mammals and birds a second locus, cytochrome b, is also widely used. Its reference library is deeper for many vertebrate groups because it was the target of phylogenetic studies before the barcoding movement formalised COI as the standard. Wildlife forensic laboratories routinely hold both primer sets and run cytochrome b when COI yields ambiguous results or when a well-characterised cytochrome b reference sequence is available for the target species.

Producing a COI sequence from a sample is technically straightforward. The reliability of the species assignment depends entirely on the quality of the reference library. BOLD Systems (barcodinglife.org) is the primary repository: it stores COI sequences linked to voucher specimens that have been identified by taxonomists, holds geographic metadata, and provides a BLAST-style identification engine that returns species assignments with percentage similarity scores.

BOLD distinguishes between sequences in its public reference library and those in private project folders. For court-ready identification the query should match a sequence in the Species Identification System: BOLD's curated, publicly accessible tier: rather than an unreviewed submission. A 98% similarity match to a voucher-verified reference sequence carries far more evidential weight than a 95% match to an anonymous GenBank submission of uncertain quality.

• Similarity thresholds: greater than 98% to the nearest BOLD record is the common acceptance criterion for species-level ID. Matches between 95-98% may indicate an unsequenced congener rather than the reference species.
• Gap statistics: some genera show no barcoding gap: intraspecific and interspecific variation overlap. Reporting should acknowledge this and offer supplementary loci or morphological corroboration where the gap is absent.
• NCBI GenBank as backup: when BOLD coverage is thin, a parallel BLAST search on GenBank provides breadth, but the provenance of reference sequences there is less tightly curated.

Short tandem repeat profiling generates a genotype at multiple hypervariable loci by measuring how many times a short motif (typically 2-6 bp) repeats at each location. Because repeat number varies independently at each locus, a multi-locus profile is statistically near-unique to an individual. The technique was developed for human forensics in the 1990s and then adapted, locus by locus, for species of commercial value in wildlife crime: African and Asian elephants, tigers, bears, and rhinoceros.

If a ranger photographs a poached elephant carcass and collects tissue before the site is abandoned, the carcass receives a reference STR profile stored in a database. When ivory is later seized, laboratory STR profiling of the tusk can be queried against the database. A match links that specific tusk to that specific carcass, establishing continuity from the crime scene to the seizure.

Many wildlife products arrive at a laboratory in a state that destroys long DNA amplicons. Dried traditional medicine preparations, heat-treated leather, steam-steamed wood, or specimens that spent months in a shipping container at high humidity all suffer DNA fragmentation. The full 648 bp COI amplicon requires intact template; when the template is shredded into fragments of 100-200 bp, the amplicon simply fails to form.

Minibarcodes solve this by targeting short sub-regions within the COI locus: typically 100 to 250 bp: that are still likely to be intact even in heavily degraded material. Several mini-barcode primer sets have been published for specific groups: vertebrates, plants, and traded invertebrates. The trade-off is reduced discriminating power, because a short fragment holds fewer informative positions. A minibarcode may resolve to family or genus when a full barcode would reach species level. The analyst's report must state which primer set was used and what taxonomic resolution it can provide.

The operational sequence in a poaching-to-seizure case looks like this. A ranger finds a carcass and collects a tissue sample: ideally from muscle or cartilage: before the scene is disturbed further. That sample is profiled at the agreed STR panel and entered into the species database. The carcass gets a unique reference number. Months or years later, a seizure of raw ivory, bone, or hide arrives at a laboratory. The seized material is profiled at the same STR panel and the genotype is queried against the database. If the profiles match at all loci and the random match probability is acceptably low, the seizure is linked to the specific carcass.

This individual-level link is more powerful than a species or population assignment because it closes a specific chain of events. Prosecutors can say: this tusk came from elephant 47-2019, which was found dead with both tusks removed in sector 3 of the reserve on 14 March 2019. The molecular evidence is the bridge between the field crime scene and the market seizure, sometimes years apart and thousands of kilometres away.

• Tissue collection at the carcass: ear notch or muscle sample placed in 95% ethanol or DMSO salt buffer. Dry storage at ambient temperature works for a few weeks; long-term requires freezing.
• Chain of custody documentation: each carcass sample must carry GPS coordinates, date, ranger ID, and continuity seal numbers: without these the database entry cannot be linked to a legally provable crime scene.
• Statistical reporting: the report should state the match probability using allele frequencies from the relevant population, not a global calculation. Elephant population structure means that allele frequencies in East Africa differ from those in Southern Africa.

Wildlife DNA evidence has reached courts in Kenya, South Africa, the United States, and several European jurisdictions, and the scrutiny applied broadly mirrors that applied to human forensic DNA. The analyst must show that the method is validated for the species, that the reference database covers the relevant populations, that the statistics are calculated correctly, and that contamination controls were adequate.

A common challenge is the claim that the database is too small to support a reliable match probability. The ElePhant database has grown steadily through contributions from national wildlife agencies and academic programmes, but it remains smaller than human forensic databases. Where the database is thin, analysts are expected to state the uncertainty explicitly rather than suppress it in a headline match figure. Courts in multiple jurisdictions have accepted wildlife DNA evidence with appropriate uncertainty statements; they have rejected it when the uncertainty was concealed.