Which cats to test next, and what each test would settle
The material on this page is drawn from the Global Cat Health Survey 2026, which is where its sources and its declarations of interest are set out in full.
Six recommendations about where to spend a test, and three about how a registry writes the trait down. None of them argues that anybody should have bred differently.
| Recommendation | Why it is that cat and not another | Where the working is |
|---|---|---|
| Test the cats whose status cannot be seen | A blue-eyed cat has already told you what it is. Its plain-eyed littermate has not, and neither has a cat its own recorder could not classify. Those are the animals where a test adds something nobody knows. | Findings |
| Trace the descendants of the unmapped lines | Six lines segregate a blue eye no laboratory has named. Their descendants are the only cohort in which those variants could be found at all. | Spread across breeds |
| Screen PAX3 in one cat of an unmapped line | All six identified determinants are PAX3 variants, and the deaf outcome reported from an Igor to Seymour cross is the signature of a compound at one locus. THIS IS THE CHEAPEST TEST ON THE LIST AND THE ONE THAT MOVES MOST: a targeted read of a single gene, not a whole-genome discovery project, and not a variant panel, which reports only the variants already named and returns clear on a seventh. | Findings |
| Test the white cats | Dominant white explains a blue eye and does not exclude a variant underneath it. A white cat is where a variant can travel entirely unseen. | Latent lines |
| A prevention control for every breed | The harm is documented in the compound heterozygote, so the control belongs at the mating and not at the individual. | Health |
| Test one cat in each cattery holding two variants in one breed group | A cattery holding two variants in breeds that are bred to one another is a place where a compound heterozygote can be made by accident rather than by intent. One result on one cat says whether the two are already in the same animals. | Variant trace |
Two variants in one cattery matter only where they sit in breeds that are bred to one another. Where a cattery holds one variant in its Maine Coons and another in its British, a compound heterozygote cannot happen by accident; a mating between them would have to be deliberate. Where both sit in the same breed group, it can.
25 catteries hold two independent variants inside one breed group, and 9 hold three or more. Across the register, 202 cats descend from a mating that held one variant on each side and could therefore carry two.
One test on one cat of such a cattery answers for the programme. It says whether the two variants are already in the same animals or still in separate lines, and that is the difference between a risk a breeder can still avoid and one already made.
The catteries and the cats are not listed here. They are computed on the pages that establish them: the catteries in which two variants could meet and every cat in which two variants could be present, both on the variant trace, and catteries in which two or more variants sit in one breed group, on the introduction.
Testing a cat that already shows the trait tells you what you can see. Every one of them points at something whose status is hidden: a plain-eyed littermate, a descendant of a line nobody has sequenced, a white cat whose coat explains the eye, a cattery in which two variants have not yet been shown to have met, or a mating rather than a cat. It is not an argument against testing the blue-eyed cat. Where a line has never been sequenced, a genotype on a cat that shows the eye is what names the variant in the first place, and nothing below can be done until it is.
These are about the record rather than about any cat. None of them asks for a test. Each asks a registry to write down what it already knows, in a form another registry can read. They are set out in full, with the figures and with the scheme applied to every record in this directory, on the proposal page.
| Recommendation | Why | Where the working is |
|---|---|---|
| Adopt a variant field, recorded against the animal | The eye codes record what an observer sees, and a latent carrier has nothing to see. A field recording which variant a cat carries, and how that was established, is the only form of the rule that reaches one. | A proposal |
| Keep 161 and give it a suffix where the line is named | 161 is live and cats are registered under it, so it is not withdrawn. Where the variant is not sequenced but the line is named, a suffix names it. A registry that ignores the suffix still reads 161, so nothing already registered becomes wrong. | A proposal |
| Gate the mating, not the cat | Carrier to carrier prohibited, carrier to unknown not permitted, carrier to clear permitted. Every documented harm in this register is a mating, and a rule written on the field reaches the lines for which no test exists. | A proposal |
This one is computed against the register on every build, so the list of animals it names moves as records are added.
A blue-eyed cat announces itself. Its brother does not. A littermate with ordinary eyes may carry the variant and pass it on unseen, and it is that animal, not the obvious one, whose descendants are now spread through the breed with nothing recorded against them.
Direct testing at the siblings of blue-eyed cats and at the animals their own recorders could not classify, rather than at the blue-eyed cats themselves, whose status is already visible. Registries hold the only records in which such animals are individually identified, and a query for the full siblings of any cat coded for this trait is a database operation rather than research. Breeders should record and code every animal in a litter, not only the blue-eyed ones: a litter recorded as a count cannot be recovered afterwards by anyone. What follows is how the directory decides which animals those are.
The directory applies a test to its own contents, computed on every build rather than typed. A cat is put forward for testing on any of three grounds, and is excluded only by a laboratory result of its own or by a source attributing its eye colour to another cause. “No blue eye” is not a finding of absence, which is the whole reason these animals need testing.
86 cats meet at least one of those three grounds, and are therefore the animals this recommendation would have tested. 18 of them have no name in any source. Two of the second group are the brothers a chart marks "Altai and Barnaul genes?", which is a suspected cross between two different unidentified variants in living, dated, pedigreed animals with green-eyed relatives available as controls. Three more are a litter whose members a compiler marked "DBE?", one of them also annotated "wide-set eyes". Five cats, and every one of them is a question the person who recorded it could not answer.
This is the question of whether a cat carries at all, and it is not the same question as which variant it carries. The cats put forward here show nothing. The cats the proposal calls the cheapest tests in the register already show the trait and are waiting to be told which variant it is. Both questions have to be answered and they are answered by different swabs, so the two lists barely overlap and neither displaces the other.
That total is a floor and a low one. It is not an estimate of the population at risk. It can only reach animals in litters entered whole, and only 70 litters covering 232 animals have been. Of the 4225 cats held here, 158 show no blue eye and 2928 have no eye colour recorded at all. Only 41 have a laboratory result of their own. Two features of the sources put a hard ceiling on any list of this kind: some litters are recorded as numbers rather than as animals, so that sixteen kittens on one chart exist only as the counts "4 kittens non-blue-eyed" and "6 kittens copper-eyed" and can never be listed or traced; and most litters are not recorded at all, because only the blue-eyed cat interested the breeder. Absence from a pedigree is not evidence that an animal did not exist or did not breed.
Testing the sibling matters because of what stands below it. A latent carrier that was never bred from matters to nobody. A latent carrier that was sold as an ordinary kitten, bred on for three generations and never recorded has put the variant into an unknown number of pedigrees in which nothing marks it. Every cat identified on any of the three grounds above should have its descendants traced forward, and a registry can run that query where no individual can. The request is specific and answerable: how many registered cats descend from this animal, in which breeds, and in which countries. It has not been asked of any registry.