Learn

DNA segments and why they matter

Shared DNA arrives in blocks, not as a smooth percentage. The size and number of those blocks tells you more about a match than the total ever will.

Updated 2026-09-13 · 8 min read · Genealogy & education only

What a segment is

DNA is not inherited as a fine powder sprinkled evenly across your genome. It arrives in blocks. During meiosis, chromosomes cross over at a handful of points, so what you receive from a parent is a set of long stretches copied from one grandparent, alternating with stretches from the other.

When a testing company says you and a match share 87 cM, what it means is: there are one or more contiguous stretches where your DNA and theirs are the same, and those stretches add up to 87 cM.

An illustration of a single chromosome pair as horizontal bars, with one matching region highlighted and labelled with its start position, end position, centimorgan length and SNP count.
A segment has a chromosome, a start, an end, a cM length and a SNP count. All five matter when you evaluate it.

Half-identical and fully identical regions

You carry two copies of every autosome, one from each parent. A half-identical region (HIR) is a stretch where at least one of your two copies matches at least one of your match's two copies. That is what almost every match is.

A fully identical region (FIR) is where both copies match on both sides. In practice this happens between full siblings, who can inherit the same piece from both parents. If a tool shows you fully identical regions with someone, full siblinghood moves sharply up the shortlist.

Why segment size matters more than count

A crossover breaks a segment. Each generation adds crossovers, so segments inherited from a distant ancestor have been chopped smaller and smaller. Longer surviving segments therefore point to a more recent common ancestor.

Longest segmentWhat it usually suggestsHow much to trust it
Over 100 cMA close, recent relationshipVery high
30–100 cMA traceable common ancestor within a few generationsHigh
15–30 cMA genuine but more distant connectionModerate
7–15 cMOften real, sometimes notCorroborate before building on it
Under 7 cMFrequently not inherited from a shared ancestor at allLow — treat as noise

Identical by descent or identical by state

Two people can match over a stretch of DNA for two very different reasons:

  • Identical by descent (IBD). You both inherited that stretch from a shared ancestor. This is the useful kind.
  • Identical by state (IBS). You happen to carry the same common variants in that region because those variants are common in the population. No shared ancestor required.

Short segments are much more likely to be IBS. That is the entire reason for a lower threshold: not because small segments are uninteresting, but because you cannot tell the two apart at that size.

A side-by-side comparison diagram: on the left, a long segment traced back through a family tree to a shared ancestor; on the right, a short segment matching two unrelated people purely by chance.
Same appearance in a match list, completely different meaning.

Sensible thresholds

  1. Ignore anything under 7 cM unless a whole cluster of evidence supports it.
  2. Check the SNP count as well as the cM length — a long segment built on very few SNPs is suspect.
  3. For a serious conclusion, want 20 cM or more, ideally triangulated with a third person.

Endogamy breaks the rules

In populations with long histories of intermarriage, everyone is related to everyone through many paths at once. Totals run high, small segments are everywhere, and the published relationship ranges become a floor rather than an estimate. If your ancestry is endogamous, lean harder on longest segment and on documentation, and lean less on total cM.

Want to look at the underlying data yourself? The Local DNA Explorer opens your raw genotype file in your browser without uploading it anywhere.