Trace ancestry refers to a very small percentage of your DNA, typically under 2%, that a genetic testing company assigns to a particular ethnic or geographic region. These tiny percentages show up on results from services like AncestryDNA and 23andMe, and they raise an immediate question: do they reflect a real ancestor from that part of the world, or are they just noise in the data? The answer is genuinely complicated, and understanding why requires knowing a bit about how DNA gets passed down and how these tests actually work.
How DNA Gets Diluted Over Generations
Every time a child is conceived, the parents’ chromosomes swap segments of DNA through a process called recombination. This shuffling means you don’t inherit a clean 50% copy from each parent. Instead, your chromosomes are a patchwork of segments from your grandparents, great-grandparents, and beyond. With each generation, the segments you inherited from any single ancestor get smaller and fewer.
The math works out roughly by halving. If one of your ancestors was 100% from a specific region, your expected share of that ancestry drops by about half with each generation. A 2% result points to an ancestor roughly five or six generations back. A 1% result suggests about seven generations, meaning a great-great-great-great-great-grandparent. For results below 1%, you’re looking at eight or more generations in the past, potentially the 1700s or earlier depending on your family’s generation length.
But “expected” is the key word. Recombination is random. You might inherit a slightly larger chunk from one distant ancestor and nothing at all from another at the same generational distance. This randomness is one reason trace percentages can shift between siblings who share the same parents. It also means that a tiny percentage could represent a real but very distant ancestor, or it could be an artifact of how the algorithm interpreted your DNA.
Why Small Percentages Are Unreliable
DNA ancestry tests work by comparing your genetic data against reference panels, which are collections of DNA samples from people with deep roots in specific regions. The algorithm looks for patterns in your DNA that statistically match patterns common in those reference groups, then assigns percentages accordingly.
Several factors make trace results especially shaky. Reference panels are incomplete. Early panels contained only around 3,000 individuals, representing just a fraction of the world’s genetic diversity. Some populations are heavily sampled while others are barely represented. When your DNA contains patterns that don’t cleanly match any reference group, the algorithm has to make its best guess, and that guess can land on the wrong region entirely.
Populations that are genetically similar to each other create additional confusion. European ancestry, for example, can be modeled as a mixture of at least three ancient populations, and the boundaries between modern European regions are genetically blurry. The algorithm might assign 1% to Scandinavia when the DNA segment in question is actually from a British ancestor whose genetics happen to overlap with Scandinavian patterns. The same problem affects other parts of the world where neighboring populations share significant genetic overlap.
At very low percentages, you’re also approaching the threshold of statistical noise. The smaller the DNA segment the algorithm is trying to classify, the less information it has to work with, and the more likely it is to produce a false match. Geneticists generally consider results below 1% to be within the range where noise becomes a serious concern.
How Algorithm Updates Change Your Results
If you’ve checked your ancestry results more than once over the years, you’ve probably noticed the numbers shifting. That’s because testing companies regularly update their algorithms and expand their reference panels. These updates can add new subregions, merge old ones, or reclassify segments of your DNA entirely.
When AncestryDNA rolled out a major update, users reported significant changes. Some saw new subregions appear and then disappear within days. Others found that entire ethnic categories were removed from their results, with the percentage redistributed to neighboring regions. In one case, a user’s Scottish percentage dropped by the same amount that their English and Northern European percentage increased, a clear sign that the algorithm was reclassifying the same DNA segments differently rather than detecting new genetic information.
These shifts are most dramatic at the trace level. A 1% result that appeared in one version of the algorithm might vanish completely in the next, not because your DNA changed, but because the company refined how it interprets certain genetic patterns. This is a strong signal that trace results should be treated as tentative rather than definitive.
What Trace Ancestry Can and Cannot Tell You
A trace result is not worthless, but it’s best understood as a hint rather than a fact. There are a few scenarios for what it might represent:
- A real distant ancestor. You may genuinely have an ancestor from that region seven or more generations back. If this aligns with family oral history or genealogical records, it strengthens the case.
- Shared ancient origins. Many populations share deep genetic roots. A small percentage assigned to one region might actually reflect ancient population movements thousands of years ago rather than a specific recent ancestor.
- Algorithm misclassification. The DNA segment might belong to a genetically similar but different population, and the reference panel wasn’t detailed enough to distinguish between them.
- Statistical noise. At percentages below roughly 1%, the result may not correspond to any real ancestry at all.
There’s no clean way to distinguish between these possibilities from the DNA test alone. The most productive approach is to cross-reference trace results with other evidence. If you have paper genealogy records, family stories, or DNA matches with people from that region, a trace result becomes more credible. Without any corroborating evidence, it’s reasonable to treat anything under 1% with healthy skepticism.
Why Results Differ Between Testing Companies
If you’ve tested with more than one company, you’ve likely noticed that trace ancestry categories don’t always agree. One service might show 1.5% West African ancestry while another shows none. This isn’t because one company is right and the other is wrong. They use different reference panels, different algorithms, and different confidence thresholds for reporting small percentages.
Some companies set a higher bar for displaying trace results, filtering out anything below a certain confidence level. Others show everything the algorithm detects, even at low confidence, and let you adjust the confidence slider yourself. 23andMe, for instance, allows you to toggle between a “speculative” mode that shows more trace categories and a “conservative” mode that collapses uncertain results into broader regional labels. The underlying DNA data is the same in both modes. What changes is how aggressively the algorithm is willing to guess.
The size and composition of the reference panel matters enormously. A company with more reference samples from West Africa will be better at detecting and correctly classifying West African ancestry than one with fewer samples from that region. As companies continue to grow their panels, trace results will likely become more accurate over time, but they will also keep shifting with each update.
Practical Ways to Interpret Your Results
If your results show a trace percentage that surprises you, the first step is to check what confidence level the estimate is set to. Many platforms default to a mid-range confidence setting. Switching to a higher confidence level will often cause trace results to disappear or fold into a broader category, which tells you the algorithm wasn’t very sure about that assignment in the first place.
Comparing results across family members can also be informative. If a sibling or parent shows the same trace ancestry at a similar or higher percentage, it’s more likely to reflect real shared heritage. If it appears in your results but not in either parent’s, that’s a strong indicator of noise or misclassification.
Building a family tree alongside your DNA results is the most reliable way to investigate. A 1% result pointing to a specific region becomes far more meaningful when you can trace an ancestor to that area in historical records. Without that paper trail, trace ancestry is an interesting data point but not something to build an identity around.