Archive | Pigment-type switching

Modifier relationships – it’s complicated


In the most recent post in this series on pigment-type switching, I talked about the fact that both horses and dogs have a wild color that is a combination of red and black pigments: bay in horses, and wolf sable in dogs. Those wild colors involve the interaction of two pigment-type switching genes, Extension and Agouti. Horses and dogs also have a mutation for all-red (located at Extension) and another for all-black (located at Agouti).

One of the most straight-forward ways to explain how this works in horses is to say that the two alleles at Extension determine if black hair is possible, and the alleles at Agouti direct the placement of the black pigment, when it is present. This flow chart might help visualize this.


So far the alleles we have here for dogs line up with those for horses. But dogs have a lot more options at both of these loci, and a look at how those work together is a great way to appreciate the complex nature of pigment-type switching, and why the chart above is only part of the picture.


I mentioned that in dogs one of the alleles at Extension, melanistic mask (EM), directs the placement of black pigment. Those familiar with horse color tend to think that is the role of the Agouti locus, which is why it is so tempting to think of Extension as the primary control, and Agouti as the modifier. However, when it comes to the alleles that control the two basic pigments, the loci themselves do not really have simple primary vs. modifier relationships. It is the individual alleles that act as modifiers, and just which allele plays that role varies with the different colors. In fact, as we’ll see later, in some cases there are multiple layers to these relationships.

So the melanistic mask allele (EM) directs the placement of the black pigment, but like horses, dogs have alleles at Agouti that also do this. In addition to the melanistic mask, the dog in the picture above has the Agouti allele for fawn (Ay), which in some breeds is called sable. Another common allele at this locus is black-and-tan (at). That is the color associated with breeds like the Doberman or Rottweiler. Both the fawn and the black-and-tan alleles direct the placement of black pigment into specific patterns.


These two alleles can work simultaneously with the allele for melanistic mask. Fawn dogs with melanistic masks are common in many breeds. This dog has both the black-and-tan pattern and the melanistic mask, which has partially obscured the tan patches that are so clearly visible on the nose of Doberman above. (This was probably even more striking before age caused the muzzle of this dog to turn gray.)


When present, black-and-tan and melanistic mask are both visible in the final coat. There are also alleles at Extension that modify specific alleles at Agouti, changing the nature of the original color. In Salukis, the color known as grizzle is an allele at Extension (EG) that modifies black-and-tan (at) . A separate allele at Extension (eh) alters black-and-tan in a similar fashion to produce “sable” English Cocker Spaniels. (The linked articles are in German, but they contain numerous images of sable (zobel) Cocker Spaniels.)

Saluki photo by Pleple2000, courtesy of Wikimedia Commons.

As the grizzle Salukis and “sable” Cocker Spaniels show, it is possible for Extension to enable the production of black pigment, direct its distribution and even modify the distribution of black pigment set down by alleles at Agouti. It is, in fact, even more complicated than this, because recent studies suggest that the black-and-tan pattern is itself a modification of a pattern known as saddle tan, which is a separate allele at a completely different locus. So grizzle is probably a modification of a modification, with the “original” pattern of red and black pigment residing at a locus that is neither Extension or Agouti.

This shows that even with dogs, where pigment-type switching has been more extensively studied, we only have a partial picture of the process. However, if we place the pieces of the puzzle that were outlined in this post into our previous flow chart, it gives a broader picture than the small window that the (known) alleles in horses allow. Here is an expanded chart showing the relationships between these alleles. If you hover over the image with your mouse, it will drop out the additional alleles to show the more limited picture provided by those alleles that correspond with the pigment-type switches in horses.

Saluki photo by Pleple2000 and Cocker photo by Louis Mayer, courtesy of Wikimedia Commons.

To make things more complicated, even this expanded flow chart covers only a portion of the pigment-type switches in dogs. While horses have two major sites that are known to control pigment-type (Extension and Agouti), dogs have a third (K Locus). That third locus, which controls dominant black and brindle, is an interesting topic for another day, but I want to take us back to horses, and what this bigger (if incomplete) picture might tell us about some of the mysteries still unsolved in that species.

Continue Reading

Wild colors and the spectrum from all-red to all-black


As I mentioned in an earlier post, the relatively small number of alleles in horses gives us a limited window into how pigment-type switches can work. This does make the basic coat colors much easier to explain, but without the bigger picture it is possible to misunderstand some fundamental aspects of animal coloration. The most common misunderstanding that I encounter is the idea that horses are “basically black” or “basically red”. Instead of Extension being one of a class of genes that control the two types of pigment (a pigment-type switch) it becomes the location where the “basic color” (red or black) is determined. This seems logical, I suspect, because there is a natural tendency to assume that anything “basic” should be reducible to one gene. So with this thinking, a horse with the dominant allele at Extension (E) becomes a “black horse”, while one with two copies of the recessive allele (ee) is a “red horse”. Any other color is the result of a “modifier”. 

But Extension is not a special site where some pure, basic color is ultimately determined. It is not the “red or black gene”—at least not in the sense that the black horse/red horse idea posits. It is part of a system where basic pigmentation (red and black) is controlled. Looking at a species that has a wider array of alleles at both Extension and Agouti can make this more obvious, which brings us to dogs. In this post I am going to outline some equivalent colors in dogs, and then follow it with a post that fills in the rest of the picture.

It is important to remember that the starting point for the color of an animal (as opposed to the color of pigment) is the wild color. In horses, that color is believed to be bay. In dogs the wild color is what is commonly called “wolf sable”. That term can be confusing, because in some breeds a genetically different color is also called sable. I have used a photo of my neighbor’s dog Caroline (above) to help highlight the difference between the two. Caroline is a German Shepherd-Rough Collie cross, and she looks quite a lot like a Collie except for her coloring. Obviously she lacks the white pattern associated with that breed, but the pattern of red (yellow) and black hair is also different. Here is my friend Melanie Miller’s Collie, Shelby, showing the other kind of sable. (Her Bully friend Barley probably also has that same gene as Shelby, but we’ll get to that later.)


If you can imagine the difference between the typical “sable” German Shepherd and a sable Collie, you can get the idea of what the wild color in dogs is. Other breeds that are commonly wolf sable are the Keeshond and the Norwegian Elkhound, but I used a picture of a German Shepherd cross because that is a familiar breed that retains the pheomelanic (red/yellow) pigment while many of the other wolf sable breeds have other factors that bleach it out, giving the impression is of a grey and black dog. This coloring is quite similar to that of the domestic dogs’ direct ancestor, the Gray Wolf. This Gray Wolf has retained a good degree of pheomelanic pigment, and looks quite similar to Caroline in color. (Photo by Carlos Delgado, and provided courtesy of Wikipedia Commons.)


Like bay in horses, wolf sable is a two-part genetic recipe. The dog needs to have the production of black pigment enabled at Extension, as well as the allele(s) for wolf sable (aw) at Agouti. As you can see, the original coloring of a dog is not red or black, nor is it something that is controlled by one primary gene. It is the product of a system of pigmentation controlled by a number of genes. Changes to that system alter the placement of the two pigments. They can also alter the proportion of the two pigments, up to and including eliminating one or the other.

Turning the animal all-red 

Dogs also have a mutation to Extension that disables the production of black pigment. Like chestnut in horses, this mutation is recessive. Dogs with this form of Extension (e) do not have black pigment in their coat, so they are all-red (yellow). This is the familiar coloring of Golden Retrievers and yellow Labradors. A recessive all-pheomelanic coloring is common in many domestic animals.


Turning the animal all-black 

Dogs also have a recessive black allele at Agouti, just like horses. The difference in dogs is that the gene for recessive black (a) is rare. It is, however, the only form of all-black possible in some breeds, like the Shetland Sheepdog. Most black dogs have a different gene, which I will discuss much later in this series. In dogs that have recessive black, though, it works like the color in horses. The dog must have black pigment enabled at Extension, and then must be homozygous for the recessive black gene (aa) at Agouti. Like recessive red, there are equivalents to this in a number of different species.


These all-red and all-black animals do not represent some essential essence of animal coloration. They are not the “basic” or “original” colors of their species. All-red (ee) and all-black (aa) are mutations that effectively turn off the production of one of the two original pigments. It can be helpful to think of them as the outermost limits for alterations to the regulation of the two pigments, with the other alleles sliding the scale to rearrange (and perhaps favor more or less) the two pigments. That arrangement can happen at either Extension or Agouti—or somewhere else entirely. I already mentioned the Melanistic Mask (EM) allele in a previous post, and how it points to the fact that Extension can direct the placement of black pigment as well as simply enable its production. I will expand on that with the next post about some of the other alleles, and hopefully give a more complete picture of what can happen with these two pigment-type switches.

Continue Reading

Translating horses to dogs


Before I return to basic dog colors, here is a short guide for translating horse color and dog color. I have posted about this a few times before, but since there is going to be a lot of new material in the next few posts, I thought a refresher on this might eliminate at least some of the potential for confusion.

Yellow = Red

In horses, we talk about two types of melanin pigment: black and red. The latter type, known as pheomelanin, is sometimes referred to as red, yellow or even orange, depending on the species. In horses, we use red. In mice, which are the most common model for research on pigmentation, it is referred to as yellow. This is also true in dogs. Because pheomelanin in horses ranges into the deeper tones—and perhaps because “yellow” was once used in some regions to mean palomino or buckskin—red is the traditional term. These are just different names for the same thing.

Red ≠ Liver or Chocolate

Which brings us to the other area of potential confusion. When horsemen talk about red, they mean pheomelanic pigment. When some dog people speak of red, what they mean is a dog like the one at the top of this post, otherwise known as liver or chocolate. On this blog, this type of coloration in dogs, which is not thought to occur in horses, is referred to by its traditional name of brown while red is reserved for pheomelanin.

Seal brown ≠ Brown (b)

Brown is the other confusing term, because it means something very different in horses than in almost any other species. In other animals, brown is an alternate form of black (eumelanin). The dog at the top of this post is the brown version of tricolor (black, tan and white). In posts on this blog, brown usually means the seal brown coloring of the sort seen on the tobiano mare below.


To avoid confusion, when speaking of the non-equine brown I follow the term with the genetic notation “(b)”. When speaking about both, I either use the genetic notations (b and At) or the more specific equine term “seal brown”. (Dogs also have a different color called seal brown, but I will leave that for another day!)

Agouti vs. agouti

In horses, as in other mammals, Agouti is a locus (a specific location on a chromosome) that is involved in pigment-type switching. In other animals, agouti is also the name given to a specific coloration that involves black and red banded hairs. Horses do not appear to have banded (agouti) hairs, but dogs do. The general convention used by this blog is that when referring to a specific locus, the name is capitalized and italicized (or underlined when in a block quote). When used to mean a pattern of banded hairs, agouti appears in regular lowercase letters. This will be an important distinction later in this conversation, because in dogs some of the alleles at Extension and Agouti produce patterns that have both agouti and non-agouti areas. So Agouti is the locus where a number of the alleles for basic coloring reside, and agouti is a specific type of coloring involving banded hairs.

I am also going to post a quick comparison of the canine equivalents – from a genetic standpoint – to bay, black and chestnut. With luck using that as a common reference point, this will not be quite as confusing.

Continue Reading