Variant trace

What each variant reaches through recorded parent links alone, with cattery prefixes ignored, recomputed on every build

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.

What this page reports is shared ancestry, not direction. Where a variant is shown to share ancestors with the Igor chart, that means the two pedigrees meet at a common ancestor. It does not mean the variant entered the Igor lines, and it does not mean the Igor line entered the variant's. Both directions have been tested and the pedigrees held exclude both: neither population is ancestral to the other. New pedigrees could change the picture in either direction.

Every variant this register holds, named and unidentified

A blue eye that is already explained is not on this page. Where a cat's own code carries a pointing factor, 31 Burmese, 32 mink or 33 colourpoint, the pointed gene is an ordinary cause of its blue eye, and where a whole breed group is of the colourpoint series it is not a group of unknown variant unless a known variant has been bred into it. One chart in this register, thirty-six cats headed Dominant Blue Eyes in Mink Ragdoll, is excluded on that ground and is counted nowhere below. The rule is set out on the method page.

The six identified variants are at PAX3 and all six are dominant, so one copy is enough to show the eye and a carrier that does not show it cannot be identified without a test. The unidentified ones are taken to be dominant on the same basis, because they produce the same phenotype in the same pattern; that is an assumption. Attributed counts every register entry to which some source gives that variant, at any evidence grade; for an unidentified variant it counts the cats of the line that show a blue or odd eye. DNA counts those with a laboratory result.

The six identified variants have been sequenced and named. The 12 unidentified ones below them are documented breeding programmes in which a dominant blue eye segregates and no animal has been genotyped. They are traced here on the same basis as the named six, since a cat carrying an unnamed allele can produce a compound heterozygote as readily as a cat carrying a named one.

DBE-CEL. The first of the six to be identified and published. The LOOF requirement of 2019 is that every Celeste descend from Lazuli, who is the one animal in this register with a laboratory result for it.

DBE-ALT. Published alongside DBE-RE, which separated the two Kazakh variants. It is the variant with the most laboratory results in this register.

DBE-RE. PAX3 c.937C>T, a nonsense variant. The cat it is named for is coded 68 on a certified pedigree, and no laboratory result for that cat himself has been published or supplied, so his own grade in this register is reported.

DBE-AGO. PAX3 c.160del, a frameshift in exon 2, published June 2025. The combination of DBE-RE and DBE-AGO is the proposed compound heterozygote risk case in the Maine Coon.

DBE-cDc. Named August 2025 and since detected in the Abyssinian within that population. It is the one variant of the six with no named cat anywhere in this register.

DBE-xHW. Named September 2025 and unpublished. It is the only one of the six with no code on any commercial panel: the laboratory that identified it sells a panel of five and this is not among them, so it is identified by exclusion and not by test.

Igor, unidentified. Four labels, one programme. Eight PawPeds sheets read 18 August 2026 put the founder in the same place with all four positions above him empty, and a second parentless coded founder, BG'Amur Dream, stands behind the Kittystar cats. Its compiler states that it reached seven named British Shorthair and Longhair catteries, SuperQueen, King Pride, Kittystar, SoloMiya, Lana's Story, Helprada and Beauty Factory, and that the founder probably carries the Altai gene, which is unconfirmed. No Altai cat is named in any account, so no pedigree can settle it. Two further statements are held and both are queried: that DBE-ALT is present in some Igor lines and it is not known which, and that cats of one cattery in this line were tested and found clear of DBE-ALT, DBE-RE and DBE-CEL. Neither names an animal, so nothing is built on either.

Blue Eyed Maine Coons (Markiss), DBE-CEL by test on the generation below. The allele this line carries is not unidentified. It is DBE-CEL, and a laboratory says so. The line begins in the Roxi by Seymour outcross into the Maine Coon, made twice over two unrelated queens, so the founding mating held one parent of each variant and either could have passed. The published genotyping on the generation below it settles which did: ten Topaz Maine Coons returned wild type for DBE-ALT and nine returned heterozygous for DBE-CEL. DBE-ALT is therefore excluded by test and DBE-CEL is carried. What has not been tested is the animals themselves. Both named offspring carry eye code 68 and neither has a result of its own, so every cat of this line is line-traced to DBE-CEL rather than genotyped for it, and it is listed here for that reason and not because the variant is unknown.

Siberian, Altai or Barnaul unconfirmed, unidentified. The largest untested population in the register by cattery count.

Nadeya, unidentified. The Ermine Trace line. Chart C13 names its variant as DBE-ALT, in red beside Oliver, beside Amelia Nanotigr and beside Ermine Trace Nadeya, who is its founder. The attribution is recorded as claimed and not adopted, because no animal in the line has been genotyped: the only laboratory result held for the founder is a colourpoint test at the C locus, which is not a variant test. The PawPeds pedigree for Angel Ermine Trace, read 18 August 2026, gives the line four complete generations above its founder, and finding 26 reports what those generations share with the Azul and Igor chart.

Topaz foundation (C6 and C11), unidentified. The founder population both Kazakh variants come out of. Neither founder has been genotyped; the variants are attributed to their lines through tested descendants.

Karagaisky, unidentified. The Munchkin line, from a village domestic cat at Chelyabinsk. Every coded cat in it also carries white spotting.

Pillowtalk-Foreigner, unidentified. The only line here that has been tested. A cat of it showed the eye and was clear of DBE-ALT, DBE-RE and DBE-CEL, which were the three variants described at the time. Retesting one cat against a current panel would either confirm a seventh variant or resolve the line to DBE-AGO. It is a German line, named for Venezia the Foreigner, and is reported restricted to the Maine Coon to date, which the register agrees with: all 61 records are Maine Coon. There is no data on homozygosity and none on deafness, and the search for the variant is reported ongoing by a German research team. Abitbol M, presentation slides, 2026, supplied to the survey. Reported: the statement carries no genotype.

Gosha, unidentified. A Russian line, its founder registered XSH with no further information recorded, and reported introduced into the Devon Rex and the Munchkin. There is no data on homozygosity and none on deafness, and the search for the variant is reported to have just started. The register holds one cat of this line and no descent from it, so neither introduction can be followed animal by animal here; the Devon Rex is carried in the breed table on the Laboratories page and no Devon Rex is entered in the register. Abitbol M, presentation slides, 2026, supplied to the survey; founder from messybeast.com. Reported: neither statement carries a genotype.

Where two variants meet, and which cats could carry both

One mating in this register brings two named variants together. On 18 July 2016, in France, Kaena, labelled DBE-ALT and recorded latent, was mated to Lazuli, DNA genotyped DBE-CEL. The litter is drawn as seven kittens on one chart and eight on another, and both readings are held. It is the only such mating the register holds.

Two variants meet in an animal only when a carrier of one is bred to a carrier of the other. A shared ancestor is a different relationship: it means two pedigrees pass through one cat, not that two alleles met in one. The shared ancestors recorded on this page were born in 2010 and 2011 and are not a meeting of alleles.

The chart labels two of the seven with both variants and those two are the two recorded deaf. Miyuki and Mochi are written DBE-CEL & DBE-ALT in red and no other kitten in the litter carries a double label. Miyuki was deaf and was not bred from. The double label is the compiler's statement of a compound heterozygote. Neither cat has been tested.

26 animals could carry two variants on this reading: the 8 of the litter and the 18 recorded below them, across 4 breeds and 2 cattery prefixes. 12 show a blue or odd eye. 26 of the 26 have never been tested for anything.

What this list is and is not. It is the population in which a compound heterozygote is possible, computed from recorded parentage. It is not a list of cats that carry two variants. Each kitten of that litter took one allele from each parent at PAX3, so each carries DBE-CEL, or DBE-ALT, or one of each, and nothing but a test distinguishes the three. Below the litter the proportion falls with every outcross.

Which DBE-CEL lines could also hold DBE-ALT

96 cats of the DBE-CEL population descend from a mating that held DBE-ALT on one side, and they fall into 6 named lines. They reach that position by more than one route: one is Kaena over Lazuli of 18 July 2016, and another is a separate outcross of Roxi into the Maine Coon, which is why the Markiss cats appear here without descending from the 2016 litter. None of the 96 has been genotyped.

A line is named here because a cat of it descends from that mating. It is not a statement that any named cat carries both variants. The lines named above are where a test would answer the question, and no cat in any of them has had one.

Every cat in which two variants could be present

Computed at every build from recorded parentage, counting both the variants the survey adopts and the ones it records a source claiming without adopting. Claimed attributions are counted here because the possibility a descendant inherits does not depend on whether the attribution met the survey's evidence threshold: Kaena's DBE-ALT is not adopted, since she was never tested, but her descendants are still in this class. 202 cats are in it.

The catteries in which two variants could meet

Two variants in one cattery are an ACCIDENTAL mating risk only where they sit in breeds that are bred to one another. Most catteries breed one breed, so in most cases they do. Where a cattery holds one variant in Maine Coons and another in British Shorthairs, a compound heterozygote cannot happen by accident: a mating between them would have to be deliberate. It would still be a prohibited mating. The rule this directory proposes bars carrier to carrier with no breed exception, because a breed boundary is a convention and the allele does not observe it. See the proposal.

25 red flagged, two independent variants inside one breed group. 9 black flagged, three or more inside one breed group. The rest hold their variants in separate breed groups.

DBE-RE and the Pillowtalk-Foreigner line are counted as two variants. A cat of the Pillowtalk-Foreigner line showed the dominant blue eye and was tested clear of DBE-RE. DBE-RE is nonetheless carried in Pillowtalk-descended pedigrees. Both statements hold if the line carries two alleles, since a line is not one genotype, and the register counts them separately on that reading.

The evidence on either side. For two alleles: the negative test above, on an animal the survey cannot name. For one: DBE-RE is named for a cat born in 2015 that descends from Pillowtalk cats, and Pillowtalk's Bon Aqua carries eye code 68 on a 2008 registration, seven years earlier. A variant named for a cat did not necessarily originate with that cat. One named Pillowtalk cat on a current six-variant panel would settle it, and would take six catteries off the flag list if the answer is one allele. Until then the register flags the pair, on the stated policy that an unconfirmed risk is recorded rather than dropped.

Where the branches part. Finding 3 gives the common ancestor: Love Hulen's Hurricane Mitch, MCO d 22, 18 January 1999. One son leads to Rociri's Elvis and so to DBE-RE; another son and a daughter lead to the Katrin Style cats and to Venezia the Foreigner. Mitch carries no eye code, and shared descent at that depth is common in this breed, so the split locates where to test rather than establishing an allele.

The limit of this flag. Finding 2 records that no cattery holding Rociri Elvis descent lacks Pillowtalk descent, so a flag on that pair fires on every DBE-RE cattery in the register and separates none of them. The table below therefore names the cats, the breed group and the route, so that the entry functions as a testing list. Naming the second variant would make the flag specific.

The breed grouping is a judgement of the survey and is set out so that it can be checked. The British group is wide because its breeds were derived from one another: the German Longhair and the Celeste both come from the British Shorthair, and the Azul is the breed of the Igor line. Maine Coon takes in the experimental longhair registrations; Siberian takes in the Neva Masquerade.

35 cattery prefixes hold cats between them covering more than one variant, and 10 of those cover three or more. A prefix is not an animal, so each row names the cats it is computed from. Rows covering three or more are outlined. They are British Gems, Dolce Blu, Emerald Flower, Ermine Trace, Hanfors world, Nikita'l, Noble Teddy House, d'Aerlin, vom Osterdeich, vom Wuerzburgerstern.

The risk is a property of the mating, not of the individual. A latent carrier of one variant bred to a carrier of another produces the compound heterozygote, and a latent carrier is not identifiable on inspection: Kaena is recorded with light green eyes and a white locket. That is the case these tables are computed to locate.

The unidentified variants

How they are named here. A variant has no official name until it is confirmed by genotype. Each row below is therefore written under the full name of the line it is possibly associated with, and nothing else: Pillowtalk-Foreigner, Igor, Karagaisky, Nadeya. The name is a label for a line, not a designation for an allele.

What this table is not. It is not a claim that these are distinct alleles. Two rows could be the same variant, or one could be a variant already named. The grouping follows the register's own line labels and nothing else, merged only where those labels are one programme written several ways. The DNA column is empty down its whole length.

The blue eyes no variant accounts for

1026 cats in this register show a blue or an odd eye. 85 of them carry a named variant. The other 941 do not, and they divide in two.

334 stand on a named line. Of those, 228 stand on a line for which no variant has been identified at all: there are 33 such lines, holding 1054 records between them, and not one animal in any of them has been genotyped. Each is a documented breeding programme with founders, dates and cattery prefixes, and no allele named.

607 more stand on no line at all, having neither traced descent nor a test. This is the largest single class in the register, and 569 of the unaccounted-for are Maine Coon.

One of these lines has been tested and came back negative. A cat of the Pillowtalk-Foreigner line showed the dominant blue eye and was clear of DBE-ALT, DBE-RE and DBE-CEL. THAT PANEL WAS THE THREE VARIANTS DESCRIBED AT THE TIME. DBE-AGO, DBE-cDc and DBE-xHW were named afterwards, so the result reads as not one of the three then known and not as a seventh variant. DBE-AGO in particular is a later-described Maine Coon variant the panel could not have covered. Retesting one cat of that line against a current panel would either confirm a seventh variant or resolve the line to DBE-AGO.

Coverage. The six identified variants account for 85 of the 1026 cats recorded here as showing a blue or an odd eye, or 8 per cent.

Tracing them

A cattery name is not a unit of inheritance. This page ignores the prefix field entirely and traces each variant through recorded parent links and nothing else.

How a dominant allele is traced, and what each column counts

A dominant allele comes down from an ancestor. The animals that could carry it are therefore the descendants of any ancestor of a known carrier. Three tiers are computed and kept apart, because they are not equally strong. Where a variant has a laboratory result anywhere, the trace is anchored on the genotyped animals only.

Attributed

Records to which some source gives this variant, at any evidence grade. A cat is counted here whether it was DNA genotyped, fixed by a first-degree genotype, traced to a named founder, or simply asserted on a breeder's chart. It is not a count of tested cats.

DNA

Of those, how many hold a laboratory result for this animal. This is the only column that rests on a test.

Below

Recorded descendants of the anchor animals. A cat here that shows the eye must have had it from the anchor.

Above

Recorded ancestors of the anchor animals. One of them transmitted it; which one is a separate question.

Collateral

The other descendants of those ancestors. These could carry it and almost none has been tested. This is the class the testing list is drawn from.

Shares ancestors with the Igor chart

Whether any animal in the three tiers stands on the breeder-compiled chart headed line descended from Altai gene Igor. It is a statement of shared ancestry and not of direction, and nothing here says whether the shared ancestors carried a variant at all.

The one variant that shares ancestors with the Igor chart

DBE-ALT, through 118 shared ancestors: Antony, CH Artos von Quintana, CH Carajan zur Sternebucht, CH Dellary's Exoti Merlin, CH Du Blue D'Or Argentin Nuance's Golden Quantin, CH Golden-Hironimus vom Goldenen Winkel, CH Golden-Ivanhoe vom Goldenen Winkel, CH High Hesket's Silver O'Connor, CH Idefix von Konradsburg, CH It's Me Ior Van Elsdieck, CH Marvellous Silver Ramzes II, CH Silver-Filou von Wahrberg, CH Sir Eddy von Elami's, Captain Ermine Trace, Du Blue D'Or Argentin Nuance's Golden Gotto, EC Annaida's Enrico Caruso, EC Darius-Doethe V Edelfeld, EC Gold-Amaretto D'Oaxaca, EC Golden-Erasmus von Wahrberg, EC Jadwiga's Silver Don Quichotte, EC Maurice vom Wahrberg, EC Quandili Quicks De Tuatha De Danann, EC Silver Valentyn van Elsdieck, Eric von Buergersruh, given on the sheet as Eric von Bürgersruh, Extasy's Cashman of Golden Flame, Flying Magic Blue Santiago of Golden Flame, GC Pennydown Top Gunn, GEC Quentin of Mystery's Dream, GIC Andy vom Wahrberg, GIC Fernando van Ybloske-Eksolby, GIC Marvellous Golden Radetzky, GIC Silver-Gary V Quintana, GIC Sir-Xenon vom Goldenen Winkel, GIC Sulver Romeo fan Liwardia, IC Bee-Jay Golden Macho Boy of Golden Flame, IC Eddie vom Kreuzanger, IC El Bimbo of Sissinghurst, IC Freedom zur Raeuberhoehle, IC Golden-Toby von Wahrberg, IC Goldfever's Golden Mikka, IC Marvellous Masterplayer, IC Sir Quincy von Elami's, IC Sulver Cupido fan Liwardia, Insider from Sham, Jadwiga's Karate Kid, Muslin Cats Patrick Peach, Rosewood vum Rosegaard, Silver Jojo from Sunnyplace, Silver-Esteban D'Oaxaca, Sonnyboy Zur Porta Westfalica, Sweet Heartbreaker's Love Me Tender, Tjarden von Felden, WCH Bluegold Nougat D'Oaxaca, WCH Greendale Sargeant Pepper, Zinedin of Mystery's Dream, Zur Sternenbucht Golden Macavity, Beauty Anak Dari Bella, CH Belia v Haus Fasler, CH Cora Dominika Peppercats, CH Di Parseghian's Vicky, CH Du Blue D'or Argentin Nuance's Golden, CH Gloria of Snow Queen, CH Golden Uglaya vom Juwelental, CH Jadwiga's Golden Undina, CH Kandy von Romero, CH Lady Di von Wahrberg, CH Lady Uliah von Elami's, CH Nicol Snow Queen, CH Noortje van de Pronkse Schonen, CH Ondra von Ursus Minor, CH Pennydown Katie Starlight, CH Philippa vom Juwelental, CH Silver Emmely from Sunnyplace, CH Silver-Whitney v Hileira, Clarinda Zolotoi Zamok, Danaclara of Sissinghurst, Du Blue D'Or Argentin Nuance's Golden Juana, EC Alice of Golden Beauty, EC Annaida's Silver Barberella, EC Ennypenny Fan Anjulies Balloo, EC Gipsy of New Celtic Henge, EC Golden-Willow von Wahrberg, EC Wilma-point von Wahrberg, Florence Rejinald, GIC Annaida's Golden Hathly, GIC High Hesket's Silver Ivory, GIC Jadwiga's Golden Rosamunde, GIC Patty von Felden, GIC Petunia Snow Queen, GIC Silver Coralie von der Anger, GIC Sulver Querina Fan Liwardia, Gaya of Pointdreams, Golden Debby vom Rosentor, Golden Flame vom Juwelental, IC Lady Bernice vom Goldenen Winkel, IC Marvellous Miss Goldfever, IC Queen of Beauty Peppercats, IC Silver Amy-Lee from Sunnyplace, IC Silver Clara vom Bleek, IC Silver-Nadine von Wahrberg, IC Yara von Felden, J'taime Aleeshane of Golden Flame, Jadwiga's Silver Classic Touch, Janosch's Chinoa, Kaori de Maneki Neko, L'Avvenires Deanna Troi, Lady Pinkybell von Elami's, Marvellous Golden Dream, Marvellous Golden Samba, Miuchi de Maneki Neko, Muslin Cats Bonita Bice, Nahimana's Leilani, Pearl of Empire's High Hope, Pearl of Empire's Jaikie, Shirley Anak Dari Bella, Tilly von Westhausen, WCH Carina vom Kreuzanger, Za-Za of New Celtic Henge. Two of them are the halves of a single mating. Every step is a recorded parent and no cattery name is used at any point. Finding 23 sets it out animal by animal.

Two branches off one mating, four generations down each side. That is a collateral relationship and it carries no direction. The shared pair is also too young to be the origin of DBE-ALT: Zinedin of Mystery's Dream was born 7 August 2011, and the variant is named for the Altai and traces to Fyodor in Kazakhstan in 1995 or 1996. Neither animal of the shared pair has been tested, so whether either carried a variant is not recorded.

The animal it ends on is Barbariska Elisa Vetsky, blue-eyed, carrying no white, four generations below the shared pair. One PAX3 panel on her says whether the Igor blue eye is DBE-ALT or a seventh variant.

What the table shows about the rest

DBE-CEL traces upward through one animal. Its one genotyped cat, Lazuli, is drawn on every chart out of an unnamed female by Roxi. The line holds 2 ancestors and 59 collateral relatives, all of them reached through that one parent link, and Roxi is computed lineage genotyped DBE-CEL: his son carries it, a dominant allele must come from a parent, and the dam is drawn as a question mark.

A trace reports the record, not the cats. Every figure on this page is a floor computed at the depth the register currently holds, and a single missing parent link is enough to make a line read as unreachable.

DBE-RE has the deepest ancestry of any variant here and reaches nothing. 173 ancestors above its genotyped animals and 187 collateral, and not one touches the Igor chart. Depth of pedigree is not what decides whether two lines meet.

No cat in the survey holds a verified genotype test showing two identified variants. A verified genotype test means a laboratory result for a named variant, on a named animal, which the survey has seen. 29 animals here hold a genotype result, and each result names one variant and stands on a different cat. The results were not commissioned or verified by the survey: they were supplied by owners or taken from work already published. Any statement that two of them meet in one cat rests on pedigree, not on a verified genotype test.

The pedigrees say the same. Five of the six identified PAX3 variants have a carrier in this register: DBE-AGO, DBE-ALT, DBE-CEL, DBE-RE, DBE-xHW. Traced from their genotyped animals and tested as 10 pairs, no two of their pedigree webs meet at a single animal. Every meeting the survey has found between blue-eyed populations is a meeting with the Igor variant, which is unidentified and is named after the cat Igor Azure Dream, and not a meeting between two identified variants.

Traced from every attributed carrier instead of from the genotyped animals only, two pairs do meet. DBE-ALT and DBE-CEL hold ancestors in common, and so do DBE-ALT and DBE-RE. Those traces run through cats that were given a variant by a chart or by descent rather than by a laboratory, so the meeting is a meeting of attributions and not of test results. The grid in the directory sets out every pairing on that wider basis.