Repeat of an old post, but interesting given today's announcements. Be sure to scroll to the bottom:
Posted by: worktoplay
In reply to: None Date:5/1/2004 12:29:58 PM
Post #of 15833
This is interesting. From the DNAWitness.net website:
DNA Witness 2.5 - About the Test
Markers:
The DNA Witness test for BGA determination is built on 176 Single Nucleotide Polymorphisms (SNPs), pronounced “snips”. SNPs are single base changes in the same location of the genetic code found in a population. SNPs result from mutations that usually arose thousands of years ago and are stably passed down through generations. SNPs are primarily bi-allelic where one parent donates one allele and the other parent donates the second allele. At a given SNP that is a variation including “G” and “A” there are three possible genotypes, “GG”, GA”, and “AA”. There are thought to be over four million SNPs in our human genomes, so why settle on 176 for DNA Witness?
The SNPs chosen for DNA Witness were found to be Ancestry Informative Markers (AIMs), that is, they exhibited a significant difference in allele frequency among different population groups. Thousands of SNPs were tested during the development of DNA Witness until the list AIMs was narrowed down to the 176 most informative SNPs that were robust and reproducible.
An example of a well-known AIM is found in the Duffy gene and is referred to as FY-Null. At this marker the frequency of one allele is .9999 for Sub-Saharan Africans, where the allele frequency in Native Americans, East Asians, and Europeans is close to zero. This sort of stark contrast is rare but there are several markers with greater than .5 delta values in allele frequency between different populations.
Populations:
The four parental populations consist of Sub-Saharan Africans, East Asians, Native Americans, and Europeans. Ninety samples from each of these four groups were genotyped to determine the allele frequencies at each marker. Obviously there are other population groups on the planet, but we felt that these four groups gave us a good starting point.
The Sub-Saharan African parental population consisted of individuals from Nigeria and the Republic of the Congo. The European parental population was from Western Europe. The East Asian parental group was from China, Japan, and Southeast Asia. The Native American parental group was from mainland Mexico.
Statistics:
A maximum likelihood calculation is used to determine the most likely population group(s) that the individual is affiliated with and at what percentage that affiliation is. For example, a tested person may be 75% European and 25% Native American.
The results are plotted for you on a standard bar graph depicting the most likely outcome and intervals of 2x confidence around those values. We also provide you with a triangle plot that is a bit less intuitive at first glance but depicts the maximum likelihood estimate (MLE) and the 2, 5, and 10X confidence contours surrounding the MLE.
Geopolitical:
The main continental groups for which the percentages are reported can be highly informative. In most cases you can eliminate 2 or 3 of the 4 main continental groups. For instance, if the result reads “92% European and 8% East Asian”, the donor of that DNA sample is excluded from being Native American, East Asian, or Sub-Saharan African in appearance. But does this mean the person is Caucasian?
We have been assembling and will continue to assemble our geopolitical database consisting of individuals that we have tested, their BGAs, and their self reported ancestry. The admixture profile from the donor of a DNA sample at a crime scene can be compared to this database to help investigators understand what possible population groups may have that profile. Obviously, there are many smaller sub-populations in the world and individuals who share traits from two or more of the major continental populations but we are trying to include as many different examples as possible and multiples of any one example.
This database adds another level of information that may prove useful in an investigation. For instance, there are individuals in the database who self describe as American Indian but their admixture of Native American on DNA Witness is low. The individual may appear to be Caucasian but may live on a reservation. If your jurisdiction contains a reservation, you would not want to discount the fact that your suspect may be living on a reservation. Following that same logic, Middle East samples can have a high level of European admixture and therefore, you would not want to exclude people from the Middle East as suspect just because the result was high European admixture and only focus on Caucasians.
Photo DB:
One of our goals is to establish a photo database of individuals from around the world linked to their BGA. Preliminary studies have indicated that a comprehensive photo database could be extremely useful in providing an investigator with an idea of what someone with a particular BGA profile might look like. Obviously, the true donor of a sample from the crime scene will not be pictured but if the investigator has 100 photos of people with BGA’s within a close range to their crime scene sample, it gives them an idea about the level of variation among the population around that BGA. We are currently seeking organizations that can help build this database rapidly. Ideally, this organization would already be taking digital mug shots and could collect a mouth swab from the subject. If you think your organization would like to participate, please send an email to forensics_info@dnaprint.com
Later,
W2P
