Understanding DNA Relationship Predictions
You open your DNA match list and see a familiar label: second cousin, close family, or third to fourth cousin. It is natural to read that label as an answer. In reality, it is usually a prediction—a useful starting point based largely on how much autosomal DNA you and the match share.
The prediction can help guide your research, but it does not establish the exact relationship by itself. Several different family relationships can produce similar amounts of shared DNA.
Why One DNA Amount Can Fit Several Relationships
Autosomal DNA is inherited through a process that mixes DNA from one generation to the next. You receive about half of your autosomal DNA from each parent, but you do not receive equal portions from every grandparent or more distant ancestor. Siblings inherit different combinations, and the variation becomes greater across generations.
Because inheritance is random, two people who share the same genealogical relationship may share different amounts of DNA. At the same time, people with different relationships may share overlapping amounts. A match predicted as a first cousin could potentially fit another relationship with a similar expected range, such as a half aunt or uncle, great-aunt or great-uncle, great-niece or great-nephew, or another relationship at a comparable genetic distance.
Centimorgans Measure Shared DNA
Testing companies commonly report shared autosomal DNA in centimorgans, abbreviated cM. A centimorgan is a unit used to describe genetic linkage; it is not a simple measure of physical distance. In everyday genealogy, the total number of shared centimorgans helps researchers evaluate which relationships are plausible.
The number matters, but it must be interpreted alongside the number and size of shared segments, the ages of the people involved, known family structure, and documentary evidence. A relationship label should therefore be read as “one or more relationships that may fit this amount of DNA.”
Why Companies May Give Different Predictions
Testing companies use their own databases, algorithms, categories, and methods for processing DNA. They may apply different rules for identifying or filtering shared segments. As a result, the same pair of people can receive slightly different shared-DNA totals or relationship labels at different companies.
The labels may also be broad. A category such as “close family” can include multiple possibilities, while “third to fourth cousin” may represent a range of relationships rather than a literal third or fourth cousin. The prediction is designed to organize a match list, not to replace genealogical analysis.
Clues That Help Narrow the Possibilities
Begin with the shared DNA total, then add every piece of context you can verify:
Compare the ages of both people. Some relationships may be biologically possible but unlikely based on the generation gap.
Build or review family trees for both matches, looking for shared surnames, locations, couples, and migration patterns.
Identify shared matches and determine whether they cluster around a known maternal or paternal line.
Review the number and size of shared segments when the testing platform provides them.
Test strategically selected relatives when appropriate. A parent, sibling, aunt, uncle, or older-generation relative may help identify which family line produced the match.
Use records to confirm the proposed pathway from each person back to the suspected common ancestor or ancestral couple.
Special Situations Can Complicate the Estimate
Endogamy occurs when generations of people marry within the same population or community. Pedigree collapse occurs when the same ancestors appear in more than one place in a family tree. Both can cause two people to share DNA through multiple ancestral pathways, which may make them appear more closely related than a single paper-trail relationship suggests.
Multiple relationships within a family can have a similar effect. In adoption, unknown-parentage, and NPE research, the relevant family tree may also be incomplete or unavailable. These situations do not make the DNA useless, but they require careful correlation and more caution when interpreting a prediction.
A Prediction Is a Hypothesis to Test
A strong relationship conclusion usually comes from combining DNA evidence with documented family relationships. The goal is not simply to choose the relationship label that looks most likely. It is to test possible relationships against ages, generations, shared matches, family trees, historical records, and any additional DNA evidence.
Sometimes the evidence strongly supports one relationship. In other cases, several possibilities may remain. Clear research acknowledges those limits and identifies the next best step rather than promising an answer the available evidence cannot yet provide.
Practical Takeaways
Treat the testing company’s relationship label as a clue, not a conclusion.
Record the exact shared centimorgans and the testing platform.
List every plausible relationship before selecting a working hypothesis.
Use ages, generations, shared matches, trees, and records to eliminate possibilities.
Remember that endogamy, pedigree collapse, and multiple relationships may affect the total.
State conclusions at the level of confidence supported by the evidence.
Have you ever discovered that a DNA match’s actual relationship was different from the testing company’s prediction?
