German Shepherd Dog Coat Colours, genetic basis, by Les Pauling, New Zealand

The gene series listed mostly go back as far as Iljin (1932), Little (1957), Burns & Fraser (1966), but modified by Willis (1976) and then again by occasional recently reported DNA findings. The situation continues to change, and I will be grateful to be given web-addresses to research not available to me when I wrote this article. The opinions are mine.

Anyone who wants to follow the “granulation” principles that explain how pigment cells actually produce the colours should look for research by or following Russell who, in 1946, found that there are 7 different ways hair pigment granules can affect the coat colour we see:

1: Granule colour
2: Granule shape (long ellipse, oval, round, irregular)
3: Granule size
4: Number of granules per central cell (medulla)
5: Number of granules per outer cell (cortex)
6: Clumping of granules, into loose masses or into denser arrangements
7: Tendency towards or away from the centre of the hair.

● Where a gene fits on a chromosome is called its locus.
● Apart from those genes carried on the sex chromosomes (in which many of the genes on the X chromosome have no partner on the smaller Y chromosome) every normal living thing (plant or animal) has 2 copies of each gene, one on each member of a chromosome pair.

Many genes have alternative versions that are known as alleles.
If both alleles in a pair are the same, the individual is homozygous for that gene/locus.
If the alleles in a pair are different, the individual is heterozygous for that gene/locus.
When two different alleles are present, normally one will suppress the other, it will be dominant; but sometimes there will be an intermediate-blend effect, such as when crossing a red flower with a white flower results in plants whose flowers are all pink.
The code for a dominant allele is always written with an upper-case initial.
The code for all other alleles is almost always written with a lower-case initial.
An allele that can be dominated is termed recessive, and will normally have no effect unless present as a homozygous pair.
When there are more than 2 alleles in a series, some of them have superscripts attached to their code, an example being the w on Aw for wolf-sable. If the typist has no ability to produce a superscript, a caret is placed in front of what should be a superscript, as A^w.
What you can see is a dog’s phenotype.
The particular genes a dog possesses are its genotype.
Any gene on any chromosome EXCEPT the X and Y ‘sex’ chromosomes is termed autosomal.

I’ll start with a warning to those who insist on any of the recessive off-colours, and those who attempt repeated line/in-breeding:

Genetic diversity IS important for the immune system.
Therefore breeders should seek diversity wherever it is possible.

We MUST attempt to eliminate the alleles that make our dogs look like some other breed.
We MUST attempt to eliminate the alleles that produce disorders & diseases.
We MUST attempt to eliminate the alleles that make our dogs unsuitable for their designated roles.

On those aspects we have no choice.

But elsewhere we must allow DIVERSITY.

The best way to ensure that is to have a Standard that prefers the dominant alleles (because then the recessives can supply diversity by being there even though their particular effect is not seen until that allele is present as a homozygous pair).

When it comes to the main genes affecting coats, only in the Intensity series does the GSD Standard not favour that diversity (but it nevertheless ALLOWS it):

	Agouti	Black/ Brown	albino Colour	Dilution	Extension	Intensity	hair Length	Spotting
Undesired Dominants						Int int^m
Desired	A^w	B	C	D	E^m	int	L	S
Recessives - mostly undesired	a^t a	b	c^ch c^d	d	E e		l	sⁱ s^p s^w

aa self black

But self-black GSDs are a a, allowing no variety in that locus; livers are b b, allowing no variety there; blues are d d, allowing no variety; longhairs are l l, allowing no variety. A pair of Izabella long-coats (b b + d d + l l) would ALWAYS produce long-coated Izabellas - no variety in colour & length of coat, and no variety in THREE genes series.

bb A liver dog

dd A blue pup on left

ll long coat

Were it not for the needs of the immune system, the ideal Standard would require everything to be recessives (of course, there'd be no breeds then, just identical dogs) - because as soon as you have a dog and a bitch that meet such a Standard then the only way you could lose that perfection in their descendants is by mutation (such as happened with Franka vom Phenom and with Australian Champion Aimsway Abacus at about the turn of the millennium - different mutations, and so different results, but in both cases the animal no longer breeds true to what its pedigree contains).

But the immune system is CRUCIAL. I'm not going to collect statistics, but I would expect that in any breed where there are varieties based on colour and/or coat length, the variety that requires any of the homozygous situations would be slightly more susceptible than the varieties that allow for heterozygosity; the susceptibility would apply to such as adult sarcomas, allergies, injection-site sarcoma, kennel cough, pancreas problems, renal failure, vaccination failure, wounds that become infected, and just plain "died young". If the individuals have pedigrees that are obviously heavily in-bred on a pooch that was itself heavily in-bred, I would expect the susceptibility to be marked.

Please don't think I am saying "Avoid in-breeding" - when properly used, in-breeding is the third-most valuable tool in the breeder's kit (first being family knowledge, second being selection).

An example of where it is essential is after you find a stud who corrects your bitch's low wither. You MUST then take your retention to that dog or his high-withered son or at least to a high-withered dog who is not related to her mother. You must in-bred on the allele for high withers, regardless of whether you in-breed on a living dog. You will probably need to do it one more time, before you can be sure that you have the high withers fixed in your bitch-line, but you might be lucky and get every pup in the second generation possessing high withers.

In-breeding on a living dog is merely the simplest way to get a pair of the desired allele, but what you're really needing is to in-breed on a particular allele, and if you can do that WITHOUT in-breeding on the first dog then you have a better chance of avoiding the dangers of in-breeding - the ideal is to double up on the "high wither allele" without doubling up on any other allele that doesn't have to be homozygous.

In each series the postulated alleles are placed with the dominant at the top (a dominant takes effect regardless of what other allele in that series it is partnered with), the recessive at the bottom (a recessive cannot take effect until present as a homozygous pair) When there are more than 2 alleles, the intermediate alleles are placed in the order in which they take effect if no allele above them in the list is present:

A = Agouti patterns

Allele symbol	Effect
A^w	wolf sable – the main guard hairs are dark tipped but have one or more bands of tan below
a^t	saddle markings with tan contrasts – the main guard hairs are dark for their whole length, unless affected by colour-paling or black-loss
a	self-colour - will be self-black unless modified by the b or d series


A^w	a^t

In many other breeds, top of the list is Ay, which produces the yellowish-brown coloration of the Agouti rodent but that is referred to as sable in such as American Cockers and Rough Collies; it may or may not be present in the GSD (I suspect “not”) and may or may not be dominant to Aw.

The phenotype (appearance) for each Agouti genotype (pair of alleles) is:

Genotype	Phenotype
A^w A^w	homozygous wolf sable
A^w a^t	heterozygous wolf sable (but carrying saddle-marking)
A^w a	heterozygous wolf sable (but carrying self-colour)
a^t a^t	homozygous saddle-marked
a^t a	heterozygous saddle-marked (but carrying self-colour)
a a	homozygous self-colour - will be self-black unless modified by the b or d series

The bi-colour pattern (literally, just 2 colours – the saddle colour plus the tan) with dark markings pencilled in on the toes and up the hocks, that is typical of the Rottweiler, is produced by an as-yet unidentified modifier, and it remains uncertain whether the modifier operates on at at, at a or a a or can modify two or three of that trio.

Although the Agouti alleles normally produce “black and tan” colorations, remember that the black can be converted by the recessive in the B, D and E series and the tan can be bleached by the recessives in the C and Int series, if the dominant in the relevant series is not present.

bi-colour

A note on sables: All well-pigmented pups are born almost fully black, with colour on just feet, cheeks, and by the anus. Self-blacks lack those colour spots. Sables (gray or gold) when dry usually show a biscuit coat with a dark stripe above the spine; at 3-4 months old a sable will be an ash colour or a dark-honey colour, and then turns dark as the sable-striped guard hairs emerge. But a true “black sable” will still be jet black at 3 months old.

A saddle-marked pup can NOT become a sable adult, but colour paling can make it look somewhat like a sable.

B = Black/Brown markings

Allele symbol	Effect
B	permits black pigment to be formed wherever the agouti pattern sets it
B	produces brown (chocolate or liver) wherever black should be, including on the “leathers” (pads, lips, nose, eye-rims).

The phenotype (appearance) for each B-series genotype (pair of alleles) is:

Genotype	Phenotype
B B	Homozygous black
B b	Heterozygous black (but carrying liver-brown)
b b	Homozygous liver-brown

C = albino Colour

Allele symbol	Effect
C	“Colour” – allows melanin to be formed
c^ch	chinchilla aka partial albinism
c^d	postulated as producing self-whites with black “leathers”
c^a	complete albinism – pink eyes & “leathers”, pale skin & hairs

Although the cd is a very convenient concept to explain the dark-nosed self-whites, there appears to be no actual evidence for it; if it exists it must be very closely associated with the e allele of the Extension series.

Regardless of whether it fits as cd or in some other series, the factor for self-white is definitely recessive; it is also epistatic, meaning that it prevents the genes for other colorations from taking effect. Many fanciers of self-whites call this the “masking” of a pattern or colour, but that can cause confusion with the black masks produced by Em.

Genotype	Phenotype
C C	Homozygous Colour, with full expression of the dark and of the tan
C c^ch	Heterozygous Colour (but carrying chinchilla); possibly the gold is lightened
C c^d	Heterozygous Colour (but carrying self-white)
c^ch c^ch	Homozygous chinchilla = partial albinism as with Norwegian Elkhounds
c^ch c^d	Heterozygous chinchilla (but carrying self-white) = partial albinism
c^d c^d	Homozygous self white

(The light gray background is to remind you that those combinations probably don’t exist.)

D = Dilution

Allele symbol	Effect
D	Dark - allows full expression of the black pigment
D	dilute – dilutes the black to a gray (aka “blue”) in both coat and “leathers”

The phenotype (appearance) for each Dilution genotype (pair of alleles) is:

Genotype	Phenotype
D D	homozygous Dark
D d	heterozygous Dark (but carrying “blue”)
d d	homozygous dilute = “blue”

A reminder here that the pairings that produce liver or blue or Izabella or self-white do not of themselves cause health problems in the GSD. In some breeds there are damaging alleles located right next to d on the chromosome, but any health problems found in liver or blue or Izabella or self-white GSDs will be due to the policies of the breeders who ignore the GSD Standard to deliberately produce these “off-colours” – the pool of partners that will produce the colour they fancy is much reduced compared to the total pool of GSDs, and so the temptation is to “forgive” whatever health problem is present in a producer of “that colour”.

When the president of the GSD Club of America wrote (Hadsell, 1972) that white “is physical evidence of genetic ingredients that will lead to degeneration in subsequent generations if not already fully demonstrated in such an individual” he might have been correct – but not for what tends to be called “genetic” reasons, just for “fanatical blindness” reasons.

E = Extension

Allele symbol	Effect
E^m	Extends dark pigment throughout the coat and adds the “mask” to the face
E	Extends dark pigment throughout the coat, but no facial mask
e	eliminates black (except in the eyes), converting it to the orange that breeders call “red”, or to shades of yellow

The phenotype (appearance) for each Extension genotype (pair of alleles) is:

Genotype	Phenotype
E^m E^m	homozygous Extended pigment with facial mask
E^m E	heterozygous Extended pigment with facial mask (but carrying maskless)
E^m e	heterozygous Extended pigment with facial mask (but carrying elimination of black, a heterozygous state which may serve to “redden” the tan)
E E	homozygous Extended pigment without mask
E e	heterozygous Extended pigment without mask (but carrying elimination of black)
e e	homozygous total elimination of black, changing it to “red” or yellow. Appears to be a requirement for self-whites

Int = Intensity

Allele symbol	Effect
Int	dilutes tan towards a dirty white
int^m	dilutes tan towards a pale yellow or fawn
int	no loss of intensity

The phenotype (appearance) for each Intensity genotype (pair of alleles) is:

Genotype	Phenotype
Int Int	homozygous dirty white aka silver
Int int^m	heterozygous dirty white (but carrying yellow-fawn)
Int int	heterozygous dirty white (but carrying full intensity)
int^m int^m	homozygous fawn
int^m int	heterozygous yellow-fawn (but carrying full intensity)
int int	homozygous full intensity

Int Int Ll

Iljin considered these alleles only partially dominant, so that each combination would produce its own blend-effect, and therefore there are 6 variations in the paling vs intensity produced by this series. Little appears to disbelieve that this series exists, but it certainly explains the very bleached fawns & silvers found in some parts of our breed.

L = hair Length

Allele symbol	Effect
L	hard, harsh, close-lying, short hair; the correct all-weather coat with the top-most guard hairs being 1-to-2 inches/2.5-to-5 cm in Length
l	long silky hair with reduced under-coat

The phenotype (appearance) for each hair Length genotype (pair of alleles) is:

Genotype	Phenotype
L L	homozygous correct-Length hard coat
L l	heterozygous correct-Length hard coat (but carrying long-hair); produces variations, from correct through “plush” to “langstockhaar”
l l	homozygous long, silky, open coat with featherings and reduced under-coat

S = Spotting

Allele symbol	Effect
S	Spotless self-colour; complete pigmentation
sⁱ	Irish spotting = white markings on the extremities
s^p	piebald spotting = roughly equal areas of white and colour
s^w	extreme white piebald, very little that is not white

Willis believes that there are various modifiers affecting this series; also that the chest blaze found in some GSD lines and that ranges from a barely visible spot to a “brassiere” is inherited independently from (i.e., is not part of) the Spotting series.

The theory refers to “plus/minus modifiers” that slow (+) or speed (-) the migration of the pigment cells during embryogenesis. Their effect is strong enough that, if a dog is S S but has several + modifiers, the pigment may be slowed enough to leave a few minor unpigmented spots, even though the dog does not have one of the alleles for actual spotting.

The phenotype (appearance) for each Spotting genotype (pair of alleles) is:

Genotype	Phenotype
S S	homozygous Spotless
S sⁱ	heterozygous Spotless (but carrying Irish spotting)
S s^p	heterozygous Spotless (but carrying piebald spotting)
S s^w	heterozygous Spotless (but carrying extreme white piebald)
sⁱ sⁱ	homozygous Irish spotting such as seen on Basenjis
sⁱ s^p	heterozygous Irish spotting (but carrying piebald spotting)
sⁱ s^w	heterozygous Irish spotting (but carrying extreme white piebald)
s^p s^p	homozygous Piebald spotting such as seen on Beagles
s^p s^w	heterozygous Piebald spotting (but carrying extreme white piebald)
s^w s^w	homozygous extreme White piebald, very little that is not white, as seen on Bull Terriers and Sealyhams

A small warning to those tempted to be careless with alleles from the Spotting series: There may be problems with eyesight if a white patch surrounds an eye; there is a strong risk of deafness if a white patch surrounds an ear. I have no information as to whether there is any intermediate-blending effect between alleles of the Spotting series, but if it happens it would help account for the variations in numbers and size of spots on spotted breeds.

However, in the GSD it is likely that only S exists, plus possibly sⁱ.

G = Graying

Allele symbol	Effect
G	progressive Graying after being born black, as with Kerry Blues.
g	grayless apart from genuine old-age effects

The phenotype (appearance) for each Graying genotype (pair of alleles) is:

Genotype	Phenotype
G G	homozygous Graying
G g	heterozygous Graying (but carrying grayless)
g g	Homozygous grayless, as per GSDs

K = black

Allele symbol	Effect
K	produces dominant self-blacK on such as Newfoundlands and Labradors
k^br	produces the brindled markings in tan areas of such as Greyhounds and Bostons
k	not black – allows “kolours” in tan areas, whether the tan “kolour” be tan or liver/brown or blue/gray or silver/white

The phenotype (appearance) for each blacK genotype (pair of alleles) is:

Genotype	Phenotype
K K	homozygous self-blacK
K k^br	heterozygous self-blacK (but carrying brindle)
K k	heterozygous self-blacK (but carrying “kolour allowance”)
k^br k^br	homozygous brindle
k^br k	heterozygous brindle (but carrying “kolour allowance”)
k k	homozygous “kolour allowance”, as per GSDs (it also allows the recessive Agouti self-black to be expressed)

M = Merle patterns

Allele symbol	Effect
M	Merle, the dappled black+gray+white pattern seen on some Rough Collies
m	non-merle

The phenotype (appearance) for each Merle genotype (pair of alleles) is:

Genotype	Phenotype
M M	homozygous Merle, associated with lethal factors, blindness & deafness
M m	heterozygous Merle (but carrying non-merle)
m m	homozygous non-merle, as per GSDs

T = Ticking

Allele symbol	Effect
T	Ticking = flecks & spots in white areas, as on Dalmatians and Pointers
t	no ticking

The phenotype (appearance) for each Ticking genotype (pair of alleles) is:

Genotype	Phenotype
T T	homozygous Ticking
T t	heterozygous Ticking (but carrying non-ticking)
t t	homozygous non-ticking; white & tan areas remain clear, as per GSDs

Series still being researched:

“Panda”-markings

These have some strong phenotype resemblances to the white markings on Collie breeds, so at first glance seem likely to be a mutation within the Spotting series. But there are complicating side-issues, so we are awaiting the often-promised-but-not-yet-delivered report from DNA-researcher Dr Mark Neff of UC Davis

Vitiligo

This is characterised by streaks of white hair & skin. To quote an article on research into the human version of it: “The genetics of vitiligo cannot be explained by simple Mendelian genetics; it is characterized by incomplete penetrance, multiple susceptibility loci and genetic heterogeneity.” So again we must wait.

Anyone who considers that I should link this article to web-sites that show well the particular coats that any of the above series produces is welcome to inform me at lesp@xtra.co.nz

Les P, March 2007

back

Agouti

Black/

Dilution

Genotype