Building a Family Tree from DNA: From Matches to Documented Ancestors
How DNA matches inform a traditional family tree: identifying common ancestors, clustering matches, and using ThruLines and Theory of Family Relativity tools.
Pillar guide
Use DNA tests for genealogy: how ethnicity estimates and centimorgan matches work, finding biological relatives, and where Y-DNA and mtDNA testing actually help.
A DNA test does two things for ancestry. It estimates the regions your DNA most resembles, and it identifies other people in the same database who share enough DNA with you to be related. Both of these are usually misunderstood at signup. The ethnicity estimate is a probabilistic comparison against a reference panel that the company updates every few years, which is why your “27% Scottish” can become “31% Irish” the next time the panel is revised. The matching is real and useful, but only inside the database you tested in.
This pillar is our complete guide to using DNA for genealogy. How matches work, how to read centimorgan amounts, how to find a biological parent or half-sibling, how to integrate DNA into a traditional family tree, and where Y-DNA and mitochondrial testing fit alongside the autosomal tests most people start with.
A few things worth knowing before you dig in.
Ethnicity estimates work better for some regions than others. British Isles and Western European reference panels are large and detailed; African, Asian, and many Indigenous American reference panels are coarser, and the breakdown reflects that limitation rather than your actual ancestry being fuzzy.
Centimorgans, abbreviated cM, are the actual currency of relationship prediction. Shared cM does not give a single answer to how two people are related; it usually narrows the options to a small set of possibilities. Our relationship predictor tool returns that set rather than a single guess, because pretending there is a single guess is how you end up confidently wrong about a relationship.
Autosomal tests like AncestryDNA and 23andMe sample DNA inherited from both parents and tell you about the past few hundred years. Y-DNA tests trace the strictly paternal line and reach much further back, but only along that one line; FamilyTreeDNA is the consumer company that still offers them. Mitochondrial (mtDNA) tests do the same for the strictly maternal line.
Matching only works within one company’s database. If your target relative tested at AncestryDNA and you tested at 23andMe, you will not see each other in either system. Uploading your raw data to MyHeritage, FamilyTreeDNA, or GEDmatch is the standard workaround.
A note on unexpected results. A meaningful share of people doing DNA tests for genealogy discover something they were not looking for: a previously unknown half-sibling, a non-paternity event, a closer biological relative than the family tree implied. We do not treat that as a curiosity. The articles below include practical guidance and pointers to support communities when a test surfaces something the family was not prepared for.
Start with the explainers below or jump straight to the article that matches your question.
How DNA matches inform a traditional family tree: identifying common ancestors, clustering matches, and using ThruLines and Theory of Family Relativity tools.
How DNA Painter works for chromosome mapping, why assigning shared segments to ancestors matters, and the practical workflow most genealogists follow.
A focused guide for adoptees using DNA testing to find biological family: which tests, how to handle privacy, when to contact, and where to find adoptee support.
Continental-level ethnicity estimates are reliable. Sub-regional estimates are best guesses, and accuracy varies enormously by region. Here is what to trust.
Half-siblings share around 1,750 cM on average, with a wide observed range. Here is how to confirm a half-sibling relationship and where to find support.
How to use DNA testing to find unknown biological relatives: which databases to test in, how to upload raw data, and where to turn for support.
How consumer DNA tests compare two people's genotype data, what total cM and longest segment mean, and why a match list only includes people in the same database.
A step-by-step search strategy for finding biological parents with DNA testing: which kits, raw data uploads, mirror trees, and where to find support along the way.
How mitochondrial DNA testing traces the strictly maternal line, what the mtFull Sequence at FamilyTreeDNA delivers, and when this test is worth ordering.
What an NPE means in genetic genealogy, how often DNA tests surface them, and where to find peer support after an unexpected result.
A reference chart of shared centimorgan amounts and the relationships they imply, drawn from the Shared cM Project. Includes overlap ranges and how to use them.
Triangulation confirms a shared ancestor when three or more people share the same DNA segment. Which databases support it and how to apply it correctly.
A centimorgan is a unit of genetic distance, not physical distance. Here is what a cM is, why total shared cM predicts relationship, and how to read the numbers.
Ethnicity estimates are probabilistic comparisons against a reference panel, not a direct readout of your ancestry. Here is how the percentages are actually calculated.
A calm, practical guide for anyone whose DNA test surfaced an unknown half-sibling, a non-paternity event, or a closer biological relative than the family tree implied.
How to upload raw DNA data to GEDmatch, what the free tools cover, when Tier 1 is worth paying for, and the law-enforcement opt-in to read first.
Haplogroups are deep-ancestry lineage labels assigned from Y-DNA and mtDNA. Here is what they mean, how they are named, and which tests actually report them.
Your DNA did not change. The reference panel did. Here is why AncestryDNA, 23andMe, and MyHeritage update their breakdowns every few years and what to expect.
Y-DNA tests trace the strictly paternal line for surname research. Here is how Y-STR and Y-SNP testing work, and how to join a surname project at FamilyTreeDNA.