Introduction
Knowing the geographic origin of a sample is often important in wildlife crime investigations, particularly where legislation varies between regions.  The collection and trade of an animal or plant may be permitted in certain areas, but prohibited in others.  In order to enforce this type of legislation using DNA evidence, it is necessary to demonstrate that a DNA sample is consistent with the genetic variation found in a particular geographic region and is extremely unlikely to have come from anywhere else. 

From a biological perspective, geographic origin identification is essentially a question of identifying the population that an individual belongs to.  Population identification is difficult compared to individual identification or species identification because the unit of measurement, the population, is not precisely defined.  A population may consist of a group of inter-related families or an entire sub-species distributed across a whole continent (Figure 1).  Therefore at one extreme, population identification can be achieved using DNA sequencing methods, similar to species identification, while at the other extreme, it requires a large number of variable genetic markers such as microsatellites (STRs) that are used in individual DNA profiling.  For further information on these techniques, please refer to the DNA identification section.



Population identification via DNA sequencing
Where populations have been isolated from one another over evolutionary time, differences between mitochondrial DNA sequences are likely to have arisen.  These population differences may correspond to specific geographic regions, such as islands, seas, or land masses divided by mountain ranges.  While the different populations have not formed separate species, there is often enough genetic sequence variation to identify a population, and therefore a geographic region of origin.  For example, mitochondrial DNA sequences have been used to identify Asian elephants from the island of Borneo (1) due to a characteristic sequence present in the population on Borneo, but not elsewhere.  Such research provides tools for conservation of endangered populations and also for wildlife forensic applications, providing evidence on the origin of illegally traded ivory.

Where categorical sequence differences occur between populations they provide very powerful markers for geographic origin identification, however for many species, such markers have not been identified.  The distribution of individuals within most species is continuous, enabling the flow of genetic material between geographic regions.  Separate populations may exist, but the transfer of DNA through individual migration or dispersal prevents the build up of unique genetic markers in populations.  From a forensic perspective, this prevents geographic origin from being identified with absolute certainty, as the DNA type observed in a population may also theoretically be found somewhere else.  In such instances, population identification relies on measuring the relative frequencies of a genetic marker in a population and then calculating the probability that an individual with that marker originates from that region.

Frequency-based approaches to population identification can also use mitochondrial DNA sequence data to measure genetic variation.  Characteristic sequences, referred to as haplotypes, can be identified within a species.  The frequency of these haplotypes will often vary among populations, allowing the origin of individuals to be inferred.  For example, a species may consist of two populations (1 & 2) each displaying three mitochondrial haplotypes (A, B, C) at the following frequencies: Pop 1 (A=49%, B=50%, C=1%); Pop 2 (A=1%, B=50%, C=49%).  If an individual of unknown origin carries haplotype A, then we can state that it is 49 times more likely to have come from Pop 1, than Pop 2.  However haplotype B individuals are equally likely to have come from Pop 1 or Pop 2.

Mitochondrial haplotype frequency differences can be informative indicators of geographic origin, but are not very useful for forensic applications as the resulting probabilities are often not powerful enough to provide convincing evidence in court.  Frequency approaches can provide stronger evidence when based on multiple genetic markers, such as microsatellites (STR), or single nucleotide polymorphisms (SNPs), used to create DNA profiles.

Population identification via DNA profiling
The use of individual DNA profiles to match two samples is an established technique, transferred directly from human forensic methods (see Individual identification using DNA).  In contrast, applying DNA profiles to geographic origin identification is less common and has no parallel with human DNA analysis.  Instead, the use DNA profiling in forensic genetic population identification is based on population assignment methods developed for evolutionary and conservation research.

DNA profiles consist of data from multiple genetic markers.  Each marker typically has a number of different states, or alleles.  The specific alleles found at each marker in an individual constitute its DNA profile.  Different alleles vary in frequency, so that some are very rare, others very common.  Allele frequency differences can be distributed geographically, in the same way that sequence haplotype frequencies can vary.  By comparing the alleles in an individual profile with the frequency of those alleles in different populations, it is possible to identify which population a sample is most likely to have originated from. 

There are two principal genetic markers used for constructing DNA profiles.  Microsatellites (or STRs) are the best known type of marker that form the basis of human profiling systems and have been developed for a large number of animal and plant species.  Microsatellite markers tend to have multiple alleles (ca. 3-10) and an individual profile would normally consist of alleles from 10-15 independent markers.  Single nucleotide polymorphisms (SNPs) are a newer type of marker, less well developed for most wildlife species.  They have fewer alleles (2-4) and therefore more markers (>50) are required to identify samples with equal confidence to microsatellite profiles, however SNPs do have several technical advantages.  Both types of marker are now used for geographic origin identification in wildlife crime investigation. 

A good example of geographic origin identification using microsatellite DNA profiles is work carried out on the identification of elephant ivory (2,3).  Understanding the origins of ivory may help direct anti-poaching efforts and uncover trade routes.  Microsatellite analysis has recently been used to track ivory seized in Singapore back to Zambia, by comparing the DNA profiles of the elephant ivory with those of elephants throughout southern Africa.  The alleles observed in the ivory allowed it to be matched to a specific population with a high degree of confidence.  SNP profiling has also recently been demonstrated for use as an enforcement tool for population identification (4).  Assignment of SNP profiles to winter or summer spawning populations in fresh water fish has successfully shown that these markers can be used to identify individuals to their original population. 

In all applications of DNA profiling to geographic origin identification, it is necessary to employ statistical analysis to calculate the probability of the sample belonging to a certain population.   There are several statistical approaches available for this (see 5 for a review) and care must be taken in both the selection  and application of the analytical technique.

References
1. Fernando P et al. (2003) DNA analysis indicates that Asian elephants are native to Borneo and are therefore a high priority for conservation. Public Library of Science, 1(1): 110

2. Wasser SK et al. (2004) Assigning African elephant DNA to geographic region of origin: Applications to the ivory trade. Proc. Natl. Acad. Sci. USA, 101 (41): 14847-14852

3. Wasser SK et al (2007) Using DNA to track the origin of the largest ivory seizure since the 1989 trade ban. Proc. Natl. Acad. Sci. USA

4. Schwenke et al. (2006) Forensic identification of a chinook salmon using a multilocus SNP assay. Conservation Genetics 7:983

5. Hauser L. et al. (2006) An empirical verification of population assignment methods by marking and parentage data: hatchery and wild steelhead (Oncorhynchus mykiss) in Forks Creek, Washington, USA, Mol. Ecol. 15 (11): 3157-3173