Forensic Analysis of Canine Samples
By comparing the canine DNA found on Professor Millstone and at the crime scene to that of the canines that belong to two suspects, Dr. Hedd and a disgruntled student, reasonable and justifiable claims about the individual responsible for Professor Millstone's death can be made. Polymerase chain reaction (PCR) as well as polyacrylamide (PAGE) and agarose gel electrophoresis are critical DNA analytical techniques that can be used to amplify certain regions of DNA using loci-specific primers and characterize the amplified DNA by separating fragments according to number of base pairs (size, weight), respectively. STR-PCR specifically amplifies loci containing two-to-ten nucleotide sequence repeats, also known as short tandem repeats (STR). When STR-PCR products are analyzed via gel electrophoresis, the variability in the number of repeats creates unique DNA profiles which can then be used as a means of identifying or distinguishing individuals.
DNA was extracted from the bulbous ends (nuclei degrade as hair cells are keratinized ) of canine hairs found on the victim, a saliva sample from the bite on her hand, hair and saliva samples from Dr. Hedd's female dog, and hair and saliva samples from the student's parent's male puppy. Also, isolated DNA from the hairs of a known female and male dog were used as a control to provide a standard of difference that could be used to better assess the similarity between the samples collected from the suspects and those collected at the crime scene. For the purpose of this lab, three canine-specific hypervariable nuclear loci (PEZ 2, PEZ 15, and VWF.X) were amplified via PCR and characterized with high resolution PAGE. Two sex-specific loci, SRY and CHR.X, were also amplified and visualized on an agarose gel. In addition to nuclear DNA, a ~500bp control region (WD3/WD6), not present in humans, containing varying numbers of 10-bp tandem repeats was also amplified and characterized via PAGE for each canine sample collected.
A general assessment of the bands generated on the PAGE gels for the nuclear loci seem to show an ambiguity and inconsistency in the band patterns, which clearly make it difficult to confidently derive accurate information from the gels. Looking at the polyacrylamide gel that had been loaded with the amplified mitochondrial locus of interest, the range of mtDNA fragment sizes from evidence samples ranged from 437-463 bp in size but relative to the control mtDNA fragments from known male and female dogs [Table 1], the samples collected from Millstone, the student, and Professor Hedd are similar [Figure 3] (at least within the margin of error). This infers that the canines from which the sample hairs originated are at least from the same maternal line. Of course, this nowhere near comes close to identifying the suspect of the crime, as the dogs from the two suspects come from the same breeder, making the two animals very likely to be related to each other through a common maternal line. Our findings support a constant profile independent of tissue type because all cells contain mitochondria and therefore mtDNA. All samples collected regardless of the source contain an approximately 500bp control region
Moreover, the DNA profile generated from DNA extracted from either hair or saliva cells for each of the individuals is vague and inconclusive, even when comparing gels from other laboratory sections. Samples comparing duplicate gels of the same samples, we concluded that all three DNA sources fit the same DNA profile and therefore shared the same DNA fragment sizes (Figures 1, 2). Since the suspect canines not only share the same maternal line, and therefore have identical mtDNA profiles, they also seem to share very similar nuclear DNA profiles, which suggests the possibility of the two dogs being directly related, and possibly even twins. The only truly distinguishing feature between the two suspect canines is their sex, but as Figure 4 illustrates, even the PAGE gel used to characterize the sex-specific loci is rather unreliable. The control samples very much exhibit clear, defined bands that signifiy the presence of either the CHR.X or SRY region, but the DNA extracted from the samples are confusing in that while we know the student owns a male dog and Professor Hedd a female dog, we see the presence of the SRY locus on samples collected from Hedd.
Because the electrophoresis analysis of both the mtDNA and nuclear DNA seem to imply that the suspects share not only a maternal line, but very similar genetic profiles, the sex-specific loci could have been used as the determining factor in identifying the suspect. The samples from Millstone’s body are possibly (though inconclusively) male and knowing the student owns a male dog, we can only assume the student was at the scene of Millstone’s death. Similarly, while Professor Hedd's dog is female, the sample was most likely mixed up or contaminated with another canine's DNA, resulting in a male genotyping of Hedd's dog. In addition to ambiguous and unexplainable bands, missing nuclear DNA evidence further belittles whatever authority can be exercised with these gels in trying to accurately identify a suspect. Further tests clearly need to be done in order to characterize evidence strong enough to support a verdict in court.
Reducing the chance of contamination by other canine DNA, diminishing any experimental error such as using the un-nucleated ends of hair samples to extract DNA or handling the thing polyacrylamide gels with more care are just a few of the obvious improvements that can be made to obtain better results. To facilitate the analysis of various DNA samples, perhaps putting DNA extracted from the same types of tissues and amplified for the same locus on one side would be a better approach than to put a variety of samples from different tissues and individuals on one gel. Another suggestion to ensure a more effective analysis of the gels would be to have a better grasp on the degree of similarity or difference between two DNA profiles, and to better understand what constitutes a difference. Of course, with any gel, there is an inherent flaw in determining molecular weight or size since this calculation not only relies on a standard curve which itself isn't 100% linear, but the creation of the standard curve necessitates a large margin of error and standard deviation since measurements must be made by hand. The lack of supporting evidence from laboratory error decreases the overall conviction using the gels as support. Other controls that could have aided in the experiment include mtDNA and nuclear DNA from the same species of dog the samples were obtained from. Separating samples throughout more gels would allow for an easier analysis of data as well. The likelihood that the mtDNA of a dog matches another dog by chance is 1 out of 14. The likelihood of the nuclear DNA of one dog matching another dog's by chance is 1 out of 526 or 0.198% [2, 3]. Therefore, the probability of having two random dogs with matching nuclear loci (PEZ 2, PEZ 15, VWF.X), mitochondrial control regions (W 3/6), and the same sex is approximately 0.007103% or 1 out of 14,078 dogs.
[1 ]J Forensic Sci, Sept. 2005, Vol. 50, No. 5: A Simplified Method for Mitochondrial DNA Extraction from Head Hair Shafts
 BC367 Experiment 5: "Forensic Analysis of Canine DNA." Colby College, Waterville, Me, Fall 2010.
. Eichmann, C., Berger, B., Steinlechner, M., and Parson, W. (2005) Estimating the probability of identity in a random dog population using 15 highly polymorphic canine STR markers, _Forensic Science International 151_, 37-44.