Author Topic: Accuracy of Racial Origin DNA Testing  (Read 5505 times)

Offline Galahad

  • The Living Force
  • ********
  • Posts: 1,205
Accuracy of Racial Origin DNA Testing
« on: November 29, 2006, 03:50:43 PM »
The following showed up on Israel Shamir's Togethernet list:

I know a bit about DNA tests.  The current tests for racial origin are based on the work of the paranoid schizophrenic mathematician Nash and implicitly rely on a hypothetical rate of mutation based on the evolution of monkeys or other primates.  If there is an error in this mutation rate or, more succintly, if there is no evolution, these tests would provide innacurate data.  I think that's the problem with the current tests.
It was the first time I had heard of this. Can some of you do some digging and see what you can uncover?

Thanks, Henry
The signs of the times herald a new age. It is close at hand. Will it be better, will it be even more dreadful than the present one? The precipice is within inches of the summit. Every one of us contributes in a certain measure to the course events takes. Let us be conscious of our responsibility

Offline Carpe

  • Dagobah Resident
  • *******
  • Posts: 734
Accuracy of Racial Origin DNA Testing
« Reply #1 on: November 30, 2006, 01:21:52 AM »
"paranoid schizophrenic mathematician Nash"  John Nash - a" beautiful mind" is meant?  John Nash was the author of couple of bioprograms like DNAfrag (restriction mapping of DNA and sizing of protein fragments) and Numclone (genomic library screen)
Maybe author meant application of Nash game theory in evolutionary biology like ESS (evolutionary stable strategy) by Maynard Smith to evaluate genes potentially surviving in a course of evolution...
It has to do with parsimony likehood molecular clock and calibrating estimates of the most recent common ancestor.

Authors claims sounds rather shallow to me. At least from that short passage you quote (and i can't find original post on togethernet)
I know a bit about DNA tests
and what?

Actually molecular clock could be regarded as at least to some extent "shallow" too as for instance mitochondrial genome evolution rate is faster than nuclear. And "ticking" of clock isn't steady too. Calibrating molecular clock is rather heated topic of debate.  Both parsimony and likehood methods suffer from approximation, so what genes to choose (slow or fast evolving, and what genome to work with) becomes of paramount importance to researcher (osit). Middle path of working with multiple slow-evolving genes from multiple genomes from multiple taxons (not just several as it occurs quite often and many taxons if chosen not carefully could produce "weird" phylogeny  just because species chosen for analysis were "weird" not only for one specific gene evolution pattern, but even for multiple).
In one word, one has to plan any evolutionary study very carefully and thoroughly taking into consideration pitfalls of molecular phylogenomic tree reconstruction. But it also depends, how deep into phylogenetic tree you like to go ... No one recipe fits all osit.

Conceptualizing human variation by Keita et al., Nature Genetics  36, S17 - S20 (2004)

Implications of biogeography of human populations for 'race' and medicine
Nature Genetics Perspective (01 Nov 2004)

Initial sequence of the chimpanzee genome and comparison with the human genome,
by The Chimpanzee Sequencing and Analysis Consortium*
Nature,  Vol 437 1 September 2005

A haplotype map of the human genome by The International HapMap Consortium Nature 437, 1299-1320 (27 October 2005)
[...]Studies of common diseases have fallen into two broad categories: family-based linkage studies across the entire genome, and population-based association studies of individual candidate genes. Although there have been notable successes, progress has been slow due to the inherent limitations of the methods; linkage analysis has low power except when a single locus explains a substantial fraction of disease, and association studies of one or a few candidate genes examine only a small fraction of the 'universe' of sequence variation in each patient.
A comprehensive search for genetic influences on disease would involve examining all genetic differences in a large number of affected individuals and controls. It may eventually become possible to accomplish this by complete genome resequencing. In the meantime, it is increasingly practical to systematically test common genetic variants for their role in disease; such variants explain much of the genetic diversity in our species, a consequence of the historically small size and shared ancestry of the human population.
Recent experience bears out the hypothesis that common variants have an important role in disease, with a partial list of validated examples including HLA (autoimmunity and infection)1, APOE4 (Alzheimer's disease, lipids)2, Factor VLeiden (deep vein thrombosis)3, PPARG (encoding PPAR ; type 2 diabetes)4, 5, KCNJ11 (type 2 diabetes)6, PTPN22 (rheumatoid arthritis and type 1 diabetes)7, 8, insulin (type 1 diabetes)9, CTLA4 (autoimmune thyroid disease, type 1 diabetes)10, NOD2 (inflammatory bowel disease)11, 12, complement factor H (age-related macular degeneration)13, 14, 15 and RET (Hirschsprung disease)16, 17, among many others.
Systematic studies of common genetic variants are facilitated by the fact that individuals who carry a particular SNP allele at one site often predictably carry specific alleles at other nearby variant sites. This correlation is known as linkage disequilibrium (LD); a particular combination of alleles along a chromosome is termed a haplotype.
LD exists because of the shared ancestry of contemporary chromosomes. When a new causal variant arises through mutation-whether a single nucleotide change, insertion/deletion, or structural alteration-it is initially tethered to a unique chromosome on which it occurred, marked by a distinct combination of genetic variants. Recombination and mutation subsequently act to erode this association, but do so slowly (each occurring at an average rate of about 10-8 per base pair (bp) per generation) as compared to the number of generations (typically 104 to 105) since the mutational event.
The correlations between causal mutations and the haplotypes on which they arose have long served as a tool for human genetic research: first finding association to a haplotype, and then subsequently identifying the causal mutation(s) that it carries. This was pioneered in studies of the HLA region, extended to identify causal genes for mendelian diseases (for example, cystic fibrosis18 and diastrophic dysplasia19), and most recently for complex disorders such as age-related macular degeneration13, 14, 15.
[...]The HapMap data were generated with the primary aim of guiding the design and analysis of medical genetic studies. In addition, the advent of genome-wide variation resources such as the HapMap opens a new era in population genetics, offering an unprecedented opportunity to investigate the evolutionary forces that have shaped variation in natural populations.
"Showing Who They Really Are":  Commercial Ventures in Genetic Genealogy
Paper presented at the American Anthropological Association Annual Meeting November 22, 2003

Deborah A. Bolnick
University of California, Davis
Over the past decade, a number of private companies have been established that make genetic testing available to the general public. At least 17 such companies currently exist in the United States and Britain, and additional ones can be found elsewhere in Europe. These companies provide a wide array of services and products, but the most common are genetic tests for reconstructing one's personal genealogical history.  These genetic genealogy tests will be the focus of my paper today.  After briefly describing the variety of tests available, I will discuss three specific examples in more detail to illustrate how such tests reinforce a number of folk ideologies about the structure of the human gene pool.
The recent proliferation of companies offering genetic genealogy tests both reflects and caters to the widespread fascination in Western society with DNA and genealogies.  According to the genealogy portal, more than 30 million Americans have set up websites tracing their family histories, and other surveys suggest that genealogy has become the second most popular hobby in the US.  Given that 80% of those surveyed say that it would be important to use DNA to determine their ancestry, these genetic genealogy tests clearly have the potential to influence a large number of people.
So, what exactly are these tests?  Generally, there are three different types available.  First, there are mitochondrial DNA tests, which sequence a portion of the mitochondrial genome.  Because mitochondrial DNA is only inherited from one's mother, this test can be used to trace maternal lineages.  Second, Y-chromosome tests examine 10 to 26 markers on the Y-chromosome, which is passed down from father to son and thus can be used to trace paternal lineages.  Closely related individuals share similar mitochondrial or Y-chromosomal DNA sequences, so these first two types of tests can be used to identify closely related individuals or groups of people.  This information can then be used to reconstruct maternal or paternal histories and to validate genealogical records.  In addition, because both mitochondrial DNA and Y-chromosome variation are geographically patterned, these tests can often identify the likely geographic origin of the lineage in question.  Finally, there is also one test of genome-wide markers, which I will discuss in more detail in a few minutes.  
Thus, for the $80-445 that a customer pays for their test, they receive their DNA sequence, information about the history and likely geographic origin of their lineage, a "framable certificate of ancestry", and some additional information about genetics and the analysis performed.
From the perspective of a molecular anthropologist, these tests clearly come with some benefits.  They make anthropological genetics accessible, relevant, and important in the eyes of the general public.  The companies provide some very good tutorials on basic genetics in laymen's terms, and they are generally good about clarifying what their tests both do and do not say.  For example, most note that their tests cannot tell you your race or ethnicity because there is no genetic basis for those classifications.
However, some companies do describe their tests in ways that are more problematic.  To help explain these issues, I want to now look in more detail at three specific examples.
First, Family Tree DNA's Cohanim test.  This test is based on a study published in 1998, which found a particular combination of Y-chromosome markers in 65% of the Jewish men who identify themselves as Cohanim, or descendents of the hereditary Jewish priesthood.  The Cohanim test looks to see if an individual has this particular set of Y-chromosome markers, which would indicate that he shares a paternal ancestor with that 65% of the Cohanim.  Interestingly, though, the company refers to this set of Y-chromosome markers as "the Cohanim gene".  Given that they define "gene" as "a functional piece of DNA that codes for a specific functional product", this suggests that it is the presence of this "gene" that confers priestly status, not the cultural or religious traditions.  In reality, however, these particular Y-chromosome markers are "junk DNA", or DNA that has no function.  Furthermore, 35% of the Cohanim are still Cohanim even though they do not possess these markers, and non-Cohanim Jews and non-Jewish Arabs who do possess the markers are still not Cohanim.  Thus, the Cohanim test conflates a genetically-defined unit with a culturally-defined one despite the fact that they are not precisely identical, and it ignores the importance of cultural factors in determining an individual's membership in this group.
Second, the Tribes of Britain test, offered by Oxford Ancestors.  This test looks at an individual's Y-chromosome to see if he exhibits markers that indicate Celtic, Anglo-Saxon, or Viking ancestry.  For example, one set of markers is common in Norway, where the Vikings originated, so the Tribes of Britain test uses it as a "genetic signature" for the Vikings.  When a British individual is identified as having this set of markers, he is told that he "can be confident that one of his paternal ancestors was a Norse Viking".  Celtic and Anglo-Saxon ancestry are identified in a similar manner.
There are, however, at least three problems with this approach to ancestry identification.  First, Oxford Ancestors is using the modern-day Norwegian population to represent the Norse Vikings.  This is certainly a reasonable thing to do, but it is valid if and only if the pattern of genetic variation in today's Norwegians is similar to that in the Vikings, who lived in Norway about 1000 years ago.  Some regions of the world do show genetic continuity through time, but others do not, presumably due to population movements or genetic drift.  Therefore, without looking at ancient DNA from the Vikings, we cannot know for certain what they looked like genetically.
Second, a given set of markers cannot be considered diagnostic of a given population unless they are unique to that population.  This mistake is made by a number of genetic testing companies, and is partly a consequence of using frequency-based conclusions from population studies to make inferences about individuals.  To illustrate this problem, let's return to the Viking example.  We'll now go ahead and assume that today's Norwegians are a good proxy for the Vikings.  In Figure 1, each circle represents a population, and the colors represent genetic markers.  Because the set of markers represented by light blue occurs at the highest frequency in Norway, any British individual with these markers most likely has Norse Viking ancestry.  However, these markers are also found in other European populations, albeit at lower frequencies.  This pattern reflects long-standing variation in Europe that predates the Vikings, so not all individuals with this particular set of markers necessarily have Viking ancestry.  
Third, this test ignores the high level of genetic variation found within all human populations.  In the Viking example, we can clearly see that some Norwegians do not possess the genetic markers represented by light blue.  This makes it likely that at least some of their Viking ancestors also exhibited different genetic markers.  
Thus, this test oversimplifies and misrepresents the patterns of genetic variation in human populations, and it does so in a way that suggests more congruence between cultural and genetic units than actually exists.  As a result, the Tribes of Britain test almost certainly identifies some individuals as having Celtic, Anglo-Saxon, or Viking ancestry when in fact they do not, and it fails to detect such ancestry in other individuals.  
The third example I want to discuss today is DNAPrint's AncestrybyDNA test.  This test examines about 70 markers found throughout the genome to estimate an individual's "precise ancestral proportions".  These proportions are then plotted on what DNAPrint calls a "Multidimensional Continuum of Ancestry" (Figure 2).  In this triangular plot, each vertex represents a different ancestry, and the single point in the triangle represents the best estimate for the individual's ancestral proportions.  In general, the closer the point is to a particular ancestry vertex, the more of that ancestry the individual has.  For the individual shown in Figure 2, the AncestrybyDNA test estimated him as having 15% Native American, 60% European, and 25% African ancestry.  
Of all the genetic genealogy tests available today, this test has the greatest implications for the public's understanding of the structure of the human gene pool.  In some ways, the test helps to break down traditional folk ideas about how human variation is patterned.  By highlighting the mixture of ancestries within each individual, for example, it demonstrates that we do not each belong to a single, pure "race".  And by depicting ancestry as something that lies along a continuum, the test results suggest that humans cannot be divided into a set number of discrete groups.
However, more often than not, this test is described in ways that instead reinforce folk ideologies about race and the pattern of human genetic variation.  The company uses the phrase "racial mix" interchangeably with "ancestral proportions", and they explicitly claim that this test measures "the biological or genetic component of race". In the frequently asked questions section of their website, the company responds to the question "what is race?" with this statement:  "Over the past few decades, there has been a movement in several fields of science to oversimplify the issue, declaring that race is 'merely a social construct.' While, indeed this may often be true, depending on what aspect of variation between people one is considering, it is also false for many particular instances of differences between the populations of the world.  One clear example of a biological difference is skin color."  Thus, they suggest that there is a biological basis to race, and they mention skin color differences as evidence for this.  The crucial detail that is being ignored here, of course, is the fact that there are biological differences between populations that are not racial differences.  Indeed, skin color is actually one of these.  It is clinally-distributed, and therefore does not divide humans into discrete groups or races.
Furthermore, the AncestrybyDNA test actually looks for only four ancestries in an individual - African, Indo-European, Native American, and East Asian.  These particular ancestries are targeted because, according to the company, they are "the four unique lineages or major population groups of the human population".  The company website also states that "in human populations, there are individuals of relatively pure BioGeographical Ancestry [their term for biological race], such as sub-Saharan Africans from Nigeria, Indo-Europeans from northern Europe, East Asians from northern China, and Native Americans from isolated regions of southern Mexico.  In other places, there are recently admixed peoples..."  Thus, this test and the associated literature reinforce the traditional racial view of human population history, in which humans divided long ago into four discrete groups, which then remained relatively isolated from one another until only recently.  Individuals from the most distant areas of each population, and therefore with the most distinct features, are taken to represent the original, pure types.  This version of human history ignores the existence and impact of gene flow, which has been an important force throughout the history of our species.
Given this additional context, it is worth re-examining what the AncestrybyDNA test actually tells you.  Based on customer comments about the test, it is clear the general public believes that it provides an individual's racial percentages, and that it has genetically verified the existence of these four human races.  What the test is actually doing, though, is surveying genetic markers that show drastic frequency differences between those so-called "pure" individuals.  For example, a genetic marker called the Duffy Null allele is found in virtually all sub-Saharan Africans, but not outside of Africa, so any person with this marker is likely to have some African ancestry.  The vast majority of human genetic variation does not follow this pattern, but this test emphasizes the very few markers that may do so.  I say "may" because there is nothing in the published material about this test to suggest that a more complete representation of each grouping was actually examined.  Thus, the observed differences might really be of a clinal nature, but there is no way to tell if only the extremes are compared.
I should note that the company has tried to validate the accuracy of this test by performing blind tests and then comparing the results with self-reported classifications.  In more than 98% of these blind tests, the majority ancestry estimates did match the individual's self-reported classification.  In other words, the test found mostly African ancestry in individuals who identified themselves as African-American, mostly European ancestry in European-Americans, and so on.  While this does validate the test's ability to genetically classify individuals as they would classify themselves, it does not validate the division of humans into four discrete groups upon which that classification is based.  
In fact, there are actually some aspects of the test results themselves that indicate problems with both the test and these groupings.  First, many individuals who believe they have a small percentage of Native American ancestry receive test results indicating a small percentage of East Asian ancestry instead.  This suggests that the test has trouble distinguishing between those two groups, presumably because they are not really genetically distinct. Second, the results for particular individuals have changed significantly over the past year.  Table 1 shows the ancestry proportions for the CEO of DNAPrint (Dr. Frudakis) and his family, which are provided on the company's website as an example.  Interestingly, their "precise ancestral proportions" changed drastically between last March, when I first looked at the site, and last week.  Given that no official changes have been made to the test in this period, this suggests that the test is not as precise or scientifically rigorous as it is claimed to be.  This is especially troubling because the results of these tests dramatically affect how individuals construct their identities, and they are being used to obtain race- or ethnicity-specific scholarships, commercial opportunities, government entitlements, and college admissions.
Thus, in conclusion, many of these genetic genealogy tests clearly reinforce folk ideologies about the structure of the human gene pool.  They suggest that race is biologically determined, and that humans can be divided into a small number of discrete groups.  They oversimplify and misrepresent the pattern of human genetic variation, and they suggest that genetic units are more congruent with racial and cultural ones than they actually are.  The persistence of such beliefs in association with these tests is particularly problematic, because these tests have a direct and profound influence on how people perceive themselves on how they are perceived by others.
list of publications in population genetics, regional
Alien baboon killing with some secret keys

Offline Akopirnas

  • Padawan Learner
  • ***
  • Posts: 93
Accuracy of Racial Origin DNA Testing
« Reply #2 on: December 11, 2007, 12:45:01 PM »
_  Hebrew University scientists say yes... :P