Genetic Genealogy: Using DNA Testing for Family Research
(summarized from a presentation by David Bradford to the Rock County Genealogical Society, March 15, 2005)
Return to RCGS Home Page
Return to Rock Research Strategies and Tools Page
View List of DNA Testing Labs
Genetic genealogy – or the use of DNA testing to help advance your family history research – has become increasingly popular in recent years. DNA has been called the molecule of life, since within a single human cell resides 46 chromosomes containing nine meters of tightly wound filaments of genetic material. These DNA filaments contain the genetic code for building proteins that determine hair, skin and eye color, gender, disease resistance and, even, future personality traits.
As parents’ donate genes to their children, almost all this parental DNA is shuffled, producing children whose appearance and DNA are somewhat different from parents and other siblings. This “DNA shuffling” is crucial to human evolution, because it allows the expressions of diverse traits which help our descendants survive in the face of changing environments. However, genetic genealogists are interested in the few portions of DNA that escape this scrambling and that continue unchanged from one generation to the next. These rare areas of stable DNA create a genetic legacy that passes without alteration over many generations. As we will see, DNA testing of these stable genetic markers can help genealogists establish proof of genetic kinship between individuals.
For these reasons, DNA testing has become a practical if pricey tool for family historians. The good news is that testing will become more comprehensive and affordable over time. The bad news (or rather, disappointing reality) is that DNA test results answer only a few kinds of genealogical questions. But, if your research has hit a brick wall as a result of one of these obstinate questions, DNA testing may provide a work-around to an otherwise frustrating research road block.
DNA Testing Basics:
You painlessly collect your own DNA sample by scrubbing cells from the inside of your cheek, and return a swab and a signed consent form by mail to your DNA lab.
Some Things DNA Cannot Do:
Costs depend on which test type and how many markers your DNA test measures; typically, $100 to $300, though your DNA Surname Project may help with costs.
Your results are returned to you in a few weeks as a DNA test “Report Card” consisting of a numerical value (allele) for each of the 10 to 44 DNA markers measured.
Your DNA results (also known as your “Genotype”) are also posted anonymously online. If you choose, your consent form permits or prohibits matching “cousins” from emailing you. Some labs act as an intermediary between people with matching results so that communication only occurs after mutual agreement by both test subjects.
Additional tests can often be performed later (for an additional fee) from your original sample stored at the performing DNA Laboratory.
Your DNA test results are most useful when compared to the results of others. In general, the closer two people's results are to one another, the more closely related are the individuals.
Unlike criminal DNA testing, your results are not a unique personal DNA fingerprint. But your results are a family genotype because close kin have identical results.
Three Basic Types of DNA Testing:
Unlike medical DNA testing, your results will not reveal any genetic disorders or foretell disease. Genealogical DNA testing looks at “Junk DNA” not your genes.
When comparing your DNA test results to another person's, DNA alone will not define the exact degree of kinship; think of DNA results as another contributory piece of evidence.
However, DNA test results are very well suited to absolutely disproving a genetic relationship; very different test results always mean unrelated individuals.
Y-Chromosome DNA (Y-DNA) reveals a man’s father’s father’s line. Y-DNA is conveyed virtually unchanged from father to son in the Y (male) chromosome without any maternal contribution. Only men have Y-DNA, so only men may contribute cells for Y-DNA testing. Female genealogists often persuade male relatives to contribute DNA in their behalf. Y-DNA results must be compared to another person’s Y-DNA results to help establish kinship. Quite logically, the closer are two people’s Y-DNA results, the closer is their relationship. The average mutation rate of individual Y-DNA markers is once every 500 years – or every 20 generations. This makes Y-DNA useful for many European and American genealogists, since the earliest use of surnames in Europe dates to between 500 and 800 years ago. Consequently, Y-DNA is the most genealogically useful DNA test. Several large, free, online Y-DNA databases already exist and are searchable for purposes of comparing your Y-DNA results to others’. Matches that are identical or differ by one or two marker values are considered relevant for genealogical comparisons.
There are several approaches to Ancestral DNA testing:
Mitochondrial DNA (mt-DNA) reveals a person’s mother’s mother’s line. Mitochondrial DNA is conveyed virtually unchanged from mother to children in the mother’s egg without any paternal contribution. Men and women both receive their mother’s mt-DNA, and either may donate cells for mt-DNA testing. But mt-DNA is considered a “paternal dead end” because only women pass their mt-DNA to later generations. Again, one’s mt-DNA results must be compared to another person’s mt-DNA results to help establish kinship. But, because mt-DNA mutates only once every 10,000 years (500 generations), mt-DNA is far less suited to establishing relationships in the genealogical time frame. A few smaller online mt-DNA databases exist for comparing your results to others’. Unlike Y-DNA, only exact matches of mt-DNA are considered relevant.
Ancestral DNA (DNA of Ancient Origin) reveals ancient geographic origins of pre-historical ancestors. Ancestral DNA will not help with issues of racial category or country of origin, however. Race is a human invention originally devised to describe broad differences in human physical characteristics, but there is no scientific evidence to support the notion of race. Nationality is also a recent human creation plagued with ever-changing national boundaries for which DNA results provide little insight. Ancestral DNA results are stand-alone findings and do not require comparison to others’ results to be meaningful.
Y-DNA add-on testing for Ancient Origins - This type of Ancient DNA describes the paternal line through the father’s father’s father. Results are provided as an ancient geographical category (Native American, Northern European, Southern European, Asian and African) and expressed as an alphabetical label called a Haplogroup (e.g., Haplogroup A) which are further subdivided into Haplotypes.
Genealogical DNA – Math, Science and Suspected Kin:
mt-DNA add-on testing for Ancient Origins - This type of Ancient DNA describes the maternal line through the mother’s mother’s mother. Results are provided as one of 38 known ancient clan mothers and expressed as a “Clan Name” (e.g., Helena) whose descendants survived from ancient times to today. Seven Clan Names (mt-Haplogroups) originate from areas now considered modern Europe. These concepts were originally explained by Bryan Sykes, geneticist and author of the book, “The Seven Daughters of Eve”.
Autosomal DNA testing (Chromosomes other than the X or Y sex chromosomes) for Ancient Origins - This has been pioneered by a few labs that will provide a comprehensive look at the total mix of ancestors contributing to a person’s genetic make-up. Some labs offer a basic qualitative measure (e.g., Native American: Yes or No) and others provide quantitative results (e.g., 14% Native American, 75% Northern European, etc.).
The science underlying DNA testing is complex, but thankfully, the level of knowledge required for a genealogist to interpret and use their DNA test results is not terribly difficult. The key to interpreting DNA results lies in understanding how DNA changes over time. Small, harmless DNA changes (mutations) occur at a predictable average rate as DNA is passed from generation to generation. This is true even of the most stable DNA markers used for genealogical testing.
One often asked question is, "How do we test the DNA of long-dead ancestors?" Unless you are related to the recently exhumed Jesse James or the Russian Romanoffs, your distant ancestors’ DNA is not readily available for testing. Consequently, genealogical genetic testing is done by comparing the results of two living individuals who suspect a kinship by dint of shared surname, common geography or an intersection of other research findings. When comparing DNA results of two suspected relatives, the number of differences in their two DNA “report cards” allows us to calculate how many generations probably separate them from one another.
This probability calculation is expressed as a number of generations called the “Time to Most Recent Common Ancestor” or TMRCA (most often shown as MRCA). This calculated MRCA is a probable but not an exact number of generations because the analysis relies on the law of averages. Here is an everyday example of how the law of averages works. If we measure the heights of 100 random men and find a “normal bell-shaped curve” whose average man is 5’10”, then the most probable height of a randomly selected man from the group is also 5’10”, but this “random man” may be the one man in 100 (1% chance) whose height was 4’10” or the one man in 100 (1% chance) whose height is 6’10”.
Restated in terms of your genealogy calculation, if the most probable calculated MRCA of two suspected relatives is 5 generations, then 5 is the most probable number of generations separating your two DNA test subjects. But there is a small, but still possible chance that the MRCA is 1 generation (a person and his father) or 10 generations (a person and his great x 7 grandfather). The take-away message here is: DNA testing can tell whether there is a genetic relationship, but not who or exactly how close. DNA provides only Relative Proof (pun intended).
Standard tables (like multiplication tables) exist for you to look-up a calculated MRCA value based on:
the number of DNA markers measured, and
Here is an important consideration in weighing the "relative proof" of DNA test results: the probabilities shown in standard MRCA tables are developed with the assumption that we are comparing the DNA of two randomly selected people. In other words, MRCA assumes there is no pre-existing suspicion of kinship. However, every well conceived genealogical DNA research plan compares people's DNA based on a strongly suspected relationship like a shared a surname, common ancestral place of residence or some other fact suggesting common ancestry. Such well-documented suspicions are very powerful but non-statistical evidence that must be considered in addition to the calculated probability found in the MRCA tables. This inevitably requires you to evaluate your DNA evidence in light of the quality and reliability of all your traditional "paper" evidence.
the number of differences between the two DNA report cards.
Once you have compared your own DNA results with those of your suspected relative, and you have used the number of differences between the two test reports to look-up a calculated MRCA value, you then compare this calculated MRCA to the expected MRCA of the relationship between yourself and your suspected kin.
For example, if you and a suspected relative believe you may both share the same great grandfather, then your expected MRCA is the number of total generations separating the two of you: 3 generations separating you and your great-grandfather plus 3 generations separating your suspected relative and his great-grandfather – or a total MRCA of 6. Finding the expected MRCA requires you to “count up and down the pedigree” in this way.
Comparing the calculated MRCA (from two people’s DNA results) to the expected MRCA (suspected from your traditional "paper" research) can result in one of three possible conclusions:
Identical or very close values for calculated versus expected MRCA values provides contributory (but not absolute) proof of the suspected relationship. In the presence of other traditional research evidence, this DNA finding can be the confirmatory bit of proof you have been seeking.
The process described above outlines one of the most common uses of genealogical DNA. The comparison of the DNA test results from carefully selected individuals allows you to strongly suggest (or absolutely disprove) whether a suspected relationship exists.
Somewhat similar results outside the statistical confidence interval (a probability-based range well beyond the calculated MRCA) means that some relationship exists, but that the two test subjects are either more closely or more distantly related than you originally suspected. This finding suggests that you are on the right track, but that you need to revise your suspicions about kinship to better reflect the reality of the DNA record.
Significantly different results allow you to conclude that no relationship exists. This is the one conclusion that DNA testing allows you to make with absolute, certainty, regardless of paper records. Remember, an unexpected finding of "non-relatedness" may reflect an adoption or a instance of maternal infidelity somewhere in your line.
Surname DNA Projects
Persons with a common surname (or variants thereof) have begun forming groups to share, analyze and publish the results of their collective DNA tests. By definition, Surname DNA Projects build their DNA databases, family histories and research efforts around a surname and, hence, utilize Y-DNA as their primary tool. They are often supplemented by Ancestral Origin DNA testing. Surname DNA Projects are coordinated by a volunteer Surname Project Coordinator, and the Project’s direction, confidentiality policies and ultimate success are a function of the skill, enthusiasm and scientific knowledge of the Administrator.
Surname DNA Projects are often sponsored by a DNA Testing Service, since at their core, the Surname Project and the lab share a basic objective: test more people. Herein lays an opportunity: the lab will often underwrite some or all of the cost of your initial Y-DNA test when you request assistance through your surname’s Project Coordinator. The lab also donates server space for Surname Project web-site hosting and provides assistance interpreting results. Like other online surname resources, the Surname DNA Project hosts a bulletin board of searchable postings and maintains a mailing list to Surname Project members. It is free to join – even if you never get tested. Your surname may have several Surname DNA Projects sponsored by different DNA testing laboratories.
In general, there has been agreement among labs to standardize testing methods, markers and terminology. Sometimes a minor adjustment to one or more marker values is necessary to use your DNA results from one lab to compare to another lab’s database in order to compare results.
The objectives of Surname DNA Projects are:
To widely test kinship of surname branches & variants;
Some Limitations of DNA Testing
To build an all-surname DNA database legacy;
To help researchers strengthen or fill-in weak paper trails;
To help researchers avoid pursuing false connections;
To gain insight into the Surname’s ancient migrations
There are several limitations and restrictions inherent in DNA testing:
Most useful DNA testing is like a peep-hole into your ancestry; your field of view is limited to the direct matrilineal and patrilineal branches of your pedigree. The vast majority of your ancestors reside between these branches and are invisible to Y-DNA and mt-DNA tests.
Surprising Results from DNA Testing
While DNA Testing is ideal for disproving a suspected relationship, it can only suggest a relationship rather than absolutely prove kinship. DNA won’t substitute for traditional research.
DNA testing is well suited to revealing “whether” a relationship exists, but because it is based on probability calculations, DNA results provide only broad and inexact help in identifying the “who” or “when” of the relationship. DNA will not pinpoint your ancestor, but it will provide strong contributory evidence when combined with other traditional forms of research.
An unexpected finding of DNA testing deserves comment. Occasionally, a lineage which has been long-accepted and well-documented using traditional “paper” research sources is absolutely disproved by DNA test results. Imagine a surname society member or a founder of a Society of Descendants of Mayflower Passengers finds that he or she does not actually carry the surname or trace their DNA to a famous pilgrim. These kinds of DNA findings, while infrequent, do occur as a result of:
Imperfect or misused original research sources
These factors are collectively estimated to occur at a rate of between 1% to 2% per generation. The impact over several generations is cumulative. For example, a 10 generation lineage has between an 11% and 34% possibility of a “break in the family DNA chain”. The lesson here is, if you ask the DNA question, be prepared for the possibility of a surprising answer.
Altered or assumed surname
Maternal infidelity and/or illegitimate birth
Strategies for DNA Testing
DNA cannot solve every research question. The key to deciding whether to submit your DNA for testing is to first develop answerable questions. An answerable question has several requirements:
First, does your question involve a branch of the pedigree that can be revealed by DNA testing? For example, does the question deal with your father’s father’s line (Y-DNA), your mother’s mother’s line (mt-DNA) or relate to a Native American ancestry (Ancient Origins DNA)?
CASE STUDIES IN GENETIC GENEALOGY
Second, do you have sufficient traditional “paper” research to propose a hypothesis (theory) to answer your question?
Third, do living suspected relatives exist to test this proposed answer to your question?
Finally, will these suspected relative agree to the inconvenience, cost and revelation of confidential genetic test results required for DNA testing?
The following case studies provide examples of these strategies in real-life research dilemmas.
Y-DNA Case Study - Thomas Jefferson & Sally Hemming
This case demonstrates several lessons relevant to Y-DNA testing. According to an article published in Nature Magazine, DNA tests “proved” that Thomas Jefferson had had a liaison with his slave, Sally Hemming. The article stated that “Jefferson sired the last son, Eston through that union, and probably 6 others.”
Conclusions were based on Y-DNA samples of direct male descendants of Sally’s oldest son and descendants of a Jefferson relative, since there were no surviving male descendants from Thomas Jefferson's marriage. Lesson One: when no living descendants of an ancestor are alive to be tested, climb up the pedigree until you find an even older ancestor who has a living male descendant. In this case the researchers climbed-up Jefferson’s pedigree to his paternal grandfather and then tested a living male descendant of Jefferson’s paternal uncle.
A later issue of Nature retracted the original conclusions because at least 25 male Jeffersons (presumably with identical Y-DNA) were living at the time of Sally’s conception of Eston, including eight adult males who frequently lived at Monticello. Each of them is as likely of fathering the child as the, then, 64 year old Thomas Jefferson. Additionally, there is always a possibility of later DNA mingling from the Jefferson line with the Hemmings line. Lesson Two: Don’t leap to conclusions. These results were reported by Pulitzer Prize winning author Phillip Ellis who wrote the book, American Sphinx – The Character of Thomas Jefferson. But he jumped to his conclusion; there was no such definite proof.
Like many of us who sustain the family myth of a famous ancestor, Sally’s descendants found it more appealing to entertain a link to a US president than a commoner. Lesson Three: Know what DNA can and cannot do. As a stand-alone test (even when combined with family tales), DNA is merely suggestive of proof. Better traditional documentation might have allowed this DNA test to become a final confirmatory proof of this “Thomas Jefferson/Sally Hemming Liaison” hypothesis, though The absence of matching DNA would be definite disproof – as was the case for the 6 other children of Sally Hemming, once the results were more carefully examined.
Y-DNA Case Study - Using a Surname DNA Project
Problem: My surname is Bradford and I can trace my paternal line to Asa Bradford in mid 18th Century Vermont, but a records gap stops me from proving that he is the “Mayflower” Asa Bradford who is a documented descendant of Mayflower Pilgrim, Gov. Wm. Bradford.
Hypothesis: That “My Asa” and the “Mayflower Asa” are one-in-the-same person.
Solution: Because I have several sources that demonstrate these two Asa Bradfords share a given name and surname, a specific time frame as well as a common geographical location, DNA tests would provide near absolute proof that these two Asa Bradfords were one-in-the-same person.
So I joined a Bradford Surname DNA Project in hope of finding several well-documented direct male descendants of Gov. William Bradford with whom I can compare Y-DNA test results to prove (or disprove) my hypothesis. I sought testing through my DNA Surname Project, and testing was provided at no cost for the basic 10 marker test, which I later paid to upgrade to a 37 marker test.
I subscribe to the Bradford DNA Surname Project mailing list and periodically visit the Bradford DNA Surname Project web page to see whether another paternal branch of the family has been identified that matches my own DNA marker test results. I have also consented to allow other with exact or near matches of DNA to contact me bty email.
Y-DNA Case Study - The Online Search for Previously Unknown New Cousins
Problem: You want to Search Y-DNA Databases for new and previously “Unknown Cousins”
Solution: Compare your Y-DNA results to those online at these web sites:
mt-DNA Case Study - Identifying the Remains of the Russian Royal Family
Sorenson Molecular Genealogy Foundation
Problem: In 1990, Russian scientists unearthed the charred bodies of what is believed to be the Russian Royal Family who were executed in 1917 at the time of the Russian Revolution.
Solution: With the help of a sample of mt-DNA provided by Prince Philip, who shares the same maternal line from Queen Victoria as the Russian Royal family, the DNA tests identified the bodies of the Romanovs, the Russian imperial family. The mt-DNA results of all but two of the bodies were identical. By other means, these two were found to be the family doctor and a maid.
In a related test, tissue from an old hospital biopsy of a deceased elderly woman who had claimed to be the escaped Princess Anastasia Romanov was compared to these remains. After years of statements to the contrary, she was definitively proven to be unrelated to the Romanovs.
mt-DNA Case Study - Finding an Immigrant Girl's Ancestral Family
Problem: In 1870, the destitute Irish parents of a young daughter were forced to give her up for immigration from Ireland to Australia and she has no birth certificate. Can she be traced back to one of three Irish families?
Solution: Locate and test mt-DNA of living maternal line descendants of girl and 3 families. An exact match between the girl’s family and one of the original Irish families constitutes confirmatory proof if the other traditional research sources that limited the analysis to three original families is reliable and well-documented.
mt-DNA Case Study - Are Two Women Sisters?
Problem: Are two deceased women sisters?
Solution: Using conventional records, find and then test mt-DNA of a direct living maternal descendant from each of the suspected sisters. An identical match proves a very close relationship suggesting sister-sister, mother-daughter or granddaughter-maternal granddaughter. If the traditional paper records already demonstrate a probable sister-sister relationship, then the mt-DNA results may be confirmatory.
mt-DNA Case Study - To Which Wife do these Daughters Belong?
Problem: A male ancestor had more than one wife and we cannot attribute some of the daughters to a specific mothers. Can we match daughters to mothers?
Solution: Find and test mt-DNA of maternal descendant of all the daughters; exact matches establish a maternal family lineage. If no direct female descendants survive, try “climbing the maternal pedigree” until you find an ancestor with a living direct female descendant.
Ancient Ancestor DNA Case Study - From Where Does my Ancient Mother Come?
Problem: From where do your most ancient European direct maternal ancestors (Clan Mothers) derive?
Solution: Use Oxford Ancestors “7 Daughters of Eve” mt-DNA Test to obtain a “Clan Mother” Haplogroup designation. See: Oxford Ancestors Matriline
Ancient Ancestor DNA Case Study - Proving Native American Ancestry
Problem: You have always heard that your family has a Native American ancestor “somewhere”.
Solution: If it’s your father’s father’s line OR the mother’s mother’s line use a test from: Family Tree DNA
If it you want % mix of 4 primal ancestral groups (including Native American), use a test from: Ancestry by DNA
DNA testing has become a powerful tool capable of contributing to family history research when your question is among the few types that DNA is best suited to answer. Because of the cost of DNA testing, you should first carefully develop a DNA testing strategy. The major types of testing each answer different questions: those related to the direct uninterrupted paternal or maternal lines and those of ancient ancestral origin.
Even at their best, DNA test results are shrouded in “probability” rather than “certainty”. So, DNA test results are contributory evidence and not absolute proof. As such, DNA test results should be considered just another resource to be considered, weighed and evaluated in the light of all evidence available on a topic.
Acknowledgement for these case studies goes to FamilyTreeDNA, OxfordAncestry.com, DNAPrint.com, TraceGenetics.com and AncestryByDNA.com
Return to RCGS Home Page
Return to Rock Research Strategies and Tools Page
View List of DNA Testing Labs