How to use a Punnett square for a pairing
Every carrier-risk figure you have ever seen comes out of a grid you can draw by hand in about ten seconds. Knowing how it works is what lets you check a number instead of taking it on trust.
The grid
Each parent carries two copies of the gene and passes on one of them, at random, to each offspring. Write one parent’s two copies along the top, the other parent’s down the side, and fill in each box with the pair that combination produces.
For two carriers of a recessive condition, both N/H:
| N | H | |
|---|---|---|
| N | N/N | N/H |
| H | N/H | H/H |
Four boxes, each equally likely. One box has two normal copies, two boxes have one of each, and one box has two copies of the variant. That is the familiar one in four affected, one in two carriers, one in four clear.
Change one parent and the square changes
Put a clear mate (N/N) against the same carrier:
| N | N | |
|---|---|---|
| N | N/N | N/N |
| H | N/H | N/H |
No box has two copies of the variant, so no offspring can be affected by that condition, whatever else happens. Half are carriers. This is the arithmetic behind keeping a valuable carrier in a program and pairing it only with tested-clear mates.
The square gives genotypes. Inheritance turns them into outcomes
This is the step that gets skipped. The grid tells you which pairs of copies are possible and how likely each is. It does not tell you which of them makes a sick animal.
- For a recessive condition, only the two-copy box is affected. The middle boxes are healthy carriers.
- For a dominant condition, one copy is enough, so those middle boxes are affected too. The same square that gives one in four for a recessive gives three in four for a dominant.
So before reading any square, check the mode of inheritance the lab states for that test. Our guide on reading a genetic test result covers where to find it.
One in four is a probability, not a quota
The most expensive misreading in breeding is treating the ratio as a promise about a litter. It is not. Every conception is a separate draw from the same square, and the draws do not remember each other.
Two carriers bred together give each offspring a one in four chance of being affected. A litter of eight can come out with none affected, or with four. A mare can produce three unaffected foals from the same risky pairing and the fourth is still a one in four chance, not a certainty. The ratios show up across many matings, not inside one.
That cuts both ways. A run of healthy offspring is not evidence the pairing was safe, and it is not evidence a parent is clear. Only a test is.
One condition per square
A square handles one gene. For two conditions, draw two squares and read them separately. Where the conditions sit on different genes they are inherited independently, so the chance of an offspring being affected by both is the two probabilities multiplied, which is usually a small number that is still not zero.
Where the square runs out
It needs a known genotype on both sides. An untested parent has no row to put in the grid, and a guess based on breed reputation or a healthy pedigree is not a genotype. The square also says nothing about conditions that are not single-gene, which is most of what a breeding program cares about beyond the tested list. Our guide on what a genetic test can’t tell you covers that edge.
Where AnimalTrace fits
Breeding checks (Client Pro) run this calculation for every condition both animals have lab carrier results on record for, and show which results the figure came from rather than a bare percentage. The conditions neither animal was tested for are listed as untested, because a square cannot be drawn for them at all.