This article was originally published in National Genealogy Society Magazine as “Broken Branches: Detecting Cases of Misattributed Parentage with DNA Evidence,” and is republished here with permission.
DNA testing can sometimes uncover unexpected cases of misattributed parentage along the ancestral lines of a test taker or matches. As a result, entire limbs of a proposed family tree may be broken off and may need to be replaced through research and exploration of the biological branches.
Genetic genealogy testing can and often does reveal surprises regarding test takers’ family trees. Review of close genetic cousins can sometimes disclose a case of misattributed parentage for a test taker or for matches. Determining which individual has the case of misattributed parentage in a family tree requires additional investigation. When considering this possibility, researchers should ask the following questions:
- Are there cousins with whom the test taker expected to share DNA but does not?
- Are there cousins with whom the test taker share less DNA than would be expected given their proposed relationship?
- Are there close genetic cousins with no known relationship?
- Are there proposed ancestral lines that have no representation in the test taker’s match list, where matches might be expected to appear based on geography and family sizes?
This article explores each of these questions in the context of a hypothetical scenario for the test results of an imaginary test taker and his imaginary matches.
Are There Cousins With Whom the Test Taker Expected To Share DNA But Does Not?
Occasionally, a test taker may be aware of other relatives who have performed DNA testing (or whom they believe performed DNA testing) but who are not showing up in match lists. When this failure to match relatives occurs, researchers should explore further to ensure that the test taker and these known relatives are indeed not sharing DNA with each other.
All relatives within the range of second cousins should share at least some DNA with each other. If a known sibling, first cousin, or second cousin has performed DNA testing and is not showing in a test taker’s match list, researchers should ensure that the following statements are true:
- The relative did indeed take and submit the DNA test.
- The relative performed DNA testing at the same company where the test subject also tested. Companies maintain separate databases, so if known cousins test with a different company, they will not appear as matches.
- The relative has opted into DNA matching. Some companies offer the option of performing an autosomal DNA test to obtain ethnicity estimates or other reports while opting out of DNA matching.
- The relative’s test results have completed processing. Sometimes there is a delay in a cousin showing up in the match lists of others if the test results have just recently completed processing.
- The relative is not using an alias or unidentifiable username. Sometimes the cousin may be in the test taker’s match list, but under a username the test taker does not recognize.
If all these statements are true, there may be a case of misattributed parentage for the test taker or for the known relative. To determine which individual does not descend from the proposed common ancestors, consider the matches in each individual’s genetic test results.
Imagine that John Smith took a DNA test along with his paternal cousin, Sharon. They are both proposed grandchildren of Paul and Helen Smith. When the test results complete processing, John finds that he does not share DNA with Sharon. In this scenario, there are two main possibilities:
- John is not a descendant of Paul and Helen Smith, or
- Sharon is not a descendant of Paul and Helen Smith
SCENARIO 1: If John shares DNA with other descendants or close collateral relatives of Paul and Helen Smith, while Sharon does not, then Sharon is not a biological descendant of Paul and Helen .
SCENARIO 2: If Sharon shares DNA with descendants or close collateral relatives of Paul and Helen which John does not match, then John is not a biological descendant of Paul and Helen.
It is also possible that neither of them descends from the couple but there are no other tested descendants of collateral relatives. Those scenarios require additional analysis and exploration.
Lack of sharing DNA between known relatives strongly indicates a case of misattributed parentage for individuals who are expected to be related within the range of close family to second cousins. More distant relatives, in the range of second cousins once removed to more distant relatives, may have simply inherited different portions of their shared ancestors’ DNA and may not share DNA with each other.
To determine if this is the case for more distant known relatives, determine if one or both individuals share DNA with other descendants or collateral relatives of the proposed common ancestors.
Are There Cousins With Whom the Test Taker Shares Less DNA Than Would Be Expected Given Their Proposed Relationship?
If known relatives share half the amount of DNA that would be expected given their proposed relationship, it may be possible that they are half rather than full relatives. To explore this possibility, utilize the evaluation tools through DNA Painter and the Shared cM Project to determine the likely relationship based on the amount of shared DNA. Analyze the matches shared between them to discover if their shared cousins include collateral relatives of both of their proposed common ancestors or only one of them.
Imagine a different situation where John Smith and his paternal first cousin, Sharon, perform DNA testing at the same testing company. When the results complete processing, they are found to share just 550 centimorgans with each other. According to DNA Painter’s Share cM Project 4.0 tool, this amount of shared DNA is more likely between half first cousins (about 80-90 percent probability) than it is between full first cousins (about 10-20 percent probability). In this scenario, there are three main possibilities:
- John is a descendant of Paul and Helen Smith while Sharon is a descendant of only Paul or Helen (scenario 3).
- Sharon is a descendant of Paul and Helen Smith while John is a descendant of only Paul or Helen (scenario 4).
- Both are descendants of Paul and Helen but happen to share a low amount of DNA with each other given their proposed relationship (scenario 5).
If John has genetic cousins who are related through the ancestry of both Paul and Helen, while Sharon only has genetic cousins who are related through Helen, and she shares consistently low amounts of DNA with other descendants of Paul and Helen, then Sharon’s parent was not the biological child of Paul but was the child of Helen.
If Sharon has genetic cousins who are related through the ancestry of both Paul and Helen, while John has only genetic cousins who are related through Helen, and he shares consistently low amounts of DNA with other descendants of Paul and Helen, then John’s father was not a biological son of Paul but was the son of Helen.
If both individuals have genetic cousins who are related through the ancestry of both Paul and Helen, then they are full first cousins and simply share low amounts of DNA given their proposed relationship.
Are There Close Genetic Matches With No Known Relationship?
Another hallmark of cases of misattributed parentage is the presence in the match list of close genetic cousins (those sharing more than 200 cM) for whom no known or documented relationship to the test subject can be determined based on their proposed family trees. To determine which individual might have a case of misattributed parentage in a family tree, it is useful to consider the shared matches between the test taker and the match and determine which family tree those shared matches support.
Imagine that John Smith has an unknown but close genetic cousin, Mary, sharing 600 cM, who has an extensive six-generation family tree associated with her test results. Based on the amount of DNA they share with each other, they should be related in the range of a first cousin to first cousin once removed. John could have a case of misattributed parentage in his family tree and may biologically descend from Mary’s ancestors. Mary could have a case of misattributed parentage in her family tree and could biologically descend from John’s ancestors. Alternatively, they could both have cases of misattributed parentage in their trees and may descend from a shared common ancestor unknown to either of them.
If John’s matches shared with Mary are known descendants and collateral relatives of his paternal grandparents, Paul and/or Helen Smith, Mary is also descended from Paul and/or Helen or one of their collateral relatives, and she has a case of misattributed parentage in her family tree (scenario 6). If John’s matches shared with Mary are all descended from a set of Mary’s second-great-grandparents, John is also descended from this same couple and has a case of misattributed parentage in his family tree (scenario 7).
If none of the matches shared between John and Mary have clear relationships to the proposed family trees of Mary and John, but instead they form their own cluster of known relatives descended from a completely different couple, then both John and Mary may have cases of misattributed parentage in their family trees (scenario 8).
One additional clue that can signal a case of misattributed parentage is when a test taker’s match list lacks representation of collateral relatives through a proposed ancestral line. However, researchers should exercise caution in these situations to avoid jumping to a hasty conclusion. In this case, absence of evidence is not necessarily evidence of absence. Having no matches from a particular line does not always mean that a test taker does not biologically descend from that family.
Unrepresented family lines might be composed of several generations of small families who had only one or two children, resulting in few living descendants to test in the first place (scenario 9). Underrepresented families might be composed of recent immigrants from countries, regions, and populations that are now well-sampled in the database (scenario 10). In other cases, family members from that particular line may not have performed DNA testing yet (scenario 11). In these situations (which are not mutually exclusive, it is useful to consider the family sizes, geographic origins, and other family details for the ancestral line that is missing in the test results.
If a lack of representation from a particular line accompanies one of the scenarios discussed above (no relationship to a known tested relative, lower than expected amounts of share DNA with the known relatives, or multiple genetic cousins with a documented relationship to each other but not the test subject), then a case of misattributed parentage if likely.
Even so, the best way to explore the anomaly of missing representation is to target-test relatives from that family line, in order to test the hypothesis that the lack of matches is due to misattributed parentage at some point along that ancestral line rather than low testing representation from the family in the database.
Imagine that John has many matches who are related through the ancestry of his maternal grandparents as well as many matches through the ancestry of his paternal grandmother, but he cannot identify any matches who are related through the ancestry of his paternal grandfather Paul Smith. If Paul Smith was from a family of ten children and descended from a long line of large families in colonial America, at least some matches through this proposed ancestral line would be expected.
Meanwhile, if Paul was the only child of an only child of an only child who was an immigrant from Germany, the lack of genetic cousins from that ancestral line is more likely due to small family sizes and recent immigration from an underrepresented population in the testing database. In either case, targeted testing of a documented relative could help confirm or refute the possibility of a case of misattributed parentage.
Autosomal DNA testing can sometimes yield surprising results which might suggest a case of misattributed parentage for test takers or their matches. In order to prove or disprove this possibility, determine if there are known tested relatives who the test taker does not match.
Establish whether known relatives have lower than expected amounts of shared DNA, given their proposed relationships.
Explore the identities of genetic cousins to determine if the list of matches includes relatives of proposed ancestors.
Research the identities of matches shared between a test taker and a mystery match to determine if one, the other, or both are descended from an unexpected family.
Consider the family structure, geographic origins, and other details of proposed ancestral lines when there are no matches from a particular line, and target-test to confirm or refute the hypothesis that there is no biological relationship to a certain proposed ancestor.
By following these steps, researchers can avoid too-hasty conclusions and prove the exact nature of a case of misattributed parentage.
More Genetic Genealogy Resources:
• Schedule a 45-minute consultation with a professional genetic genealogist here if you need more help determining the exact nature of misattributed parentage.