Tag Archives | KIT mutations

Where the wild things are

NotTob

In response to the previous post about the Agouti locus, a reader questioned the importance of the fact that bay was the original color of wild horses—that it “came first”, before black or chestnut. Why should that matter?

This question touches on the reason why I have come believe that the way color gets explained matters so much. I have mentioned in previous posts that equine coat color has become far more complex since I first began writing about it. (At the risk of revealing my age, that was around 1990.) There was a time when everything could be explained in terms of the colors themselves, while the technical aspects of genetics could be skipped or at least minimized. In hindsight, while some of these explanations made certain concepts easier to grasp, they could also be misleading. Let me give an example of a common misunderstanding from fifteen years ago, the explanation that was used to clarify the situation, and how that same concept—so useful then!—is somewhat problematic now.

At the time in question, the concept of dominance was difficult for a lot of people. It was not uncommon to get questions like, “Which is more dominant, grey or bay?” Or, “I know bay is dominant to chestnut, so shouldn’t palomino be recessive to bay?” These questions arose because it was not clear that a color could not be dominant or recessive to an unrelated color. My approach was to point out that colors could only have this kind of relationship with their opposites. Tobiano, for example, could not be dominant or recessive to buckskin; tobiano could only be dominant or recessive to “not-tobiano”. The opposite of a color was not a different color, but the absence of the color. This was a very clear way to get the idea across that the genes for different colors were separate things, and that each presented an independent chance for inheritance.

BuckTobi
Fifteen years ago, many did not understand that each aspect of this horse’s color—bay, cream and tobiano—involved a separate, unrelated gene.

It was a simple explanation, but behind it lurked some puzzling questions for anyone who cared to look a little closer. If tobiano was dominant to “not-tobiano”, what exactly was this “not-tobiano”? If tobiano was believed to have arisen after domestication, how on earth were those wild ancestors carrying around a gene for the absence of something that did not yet exist? The idea of “not-tobiano” worked when it came to predicting breeding outcomes, but looked at in this light it made no sense.

That is why something like the situation with bay as an ancestral color matters, because the key to understanding what is really going on with “not-tobiano” can be found there. As I mentioned in the previous post, bay (or bay dun) is the most likely ancestral, or wild, color for horses. The other two basic colors, chestnut and black, were later mutations to the two genes responsible for bay. Another word for those alleles that were already there is wild-type. The wild-type is the allele that is typical for a given population. Wild-type is the “normal” setting—the default—for a gene. “Not-tobiano” and all those other “not-colors” were really just that: the wild-type for their particular gene. In the case of things like dilutions and white patterns, the wild-type is usually just the instructions for normal pigmentation.

Shifting from a color-based approach to a gene-based approach

Looking at colors in terms of the wild-type eliminates the misunderstandings that come from thinking of the color itself as a gene. Because we often refer to colors this way—as the “tobiano gene” or the “cream gene”—it is easy to get the idea that something like the cream dilution is an additional gene that palomino, buckskins, and smoky creams have; one which non-diluted horses do not have. The cream dilution is actually a mutation that occurred to a gene, known as MATP, that all horses have. In the absence of the cream dilution, MATP is involved in the normal formation of pigment. So the wild-type for that gene gives a fully-pigmented horse.

Not knowing there is a wild-type makes it seem that the color (Cream) is the gene itself and therefor the starting point. That is why there is a tendency to assume later discoveries are “mutations of the color” rather than alleles for the same (non-mutated, wild-type) gene. So pearl, which is found in the same genetic location as cream, becomes a “mutation of cream” rather than a second, unrelated mutation of the MATP gene. But the starting point is not cream, but the wild-type at MATP. The cream mutation did not have to be present for pearl to occur; it is a mutation like cream, not necessarily a mutation to cream.

But perhaps more importantly, many colors were named before their relationships to other colors were understood. Things that once were assumed to be separate later proved to be alleles of the same gene. At one time we thought, and taught, that the opposite of tobiano was the absence of tobiano. But the “tobiano gene” is not a separate gene. Tobiano is a mutation to the KIT gene, which again is a gene that all horses have regardless of their color. Tobiano shares the KIT gene with a host of other alleles (like Sabino1, Roan and the White Spotting patterns) that have historically been thought of as unrelated. That complex situation is very difficult to explain, especially if someone’s basic understanding of the subject is color-based rather than gene-based, because the relationship between that group of colors is not visually obvious.

Unicorn
Tobiano and roan are both alleles of the KIT gene, which is why the combination does not breed true. The offspring can only get one or the other from the parent.

I know for many who have learned about horse color exclusively in terms of basic colors and their modifiers, focusing on the actual genes is a very different approach. It may seem like it adds a lot of unnecessary complexity to the subject. I certainly can appreciate that point of view, but genes and the importance of using their wild-type as a starting point is the missing piece of the puzzle for a lot of people. When that piece falls into place, color genetics—especially as it is currently understood—begins to make a lot more sense.

Continue Reading

Some new research into blue eyes and KIT mutations

BlueNAmber1
A tobiano pony with partially blue and partially gold eyes (cause unknown)

A year ago I posted expressing my doubts about the theory that blue-eyed horses with KIT mutations (tobiano, sabino and the white spotting patterns) must carry an additional mutation to account for their eye color. At the time, I did not think that there was enough evidence to merit the absolute terms that were often used regarding this theory. It is also true that, having tracked so many instances of blue eyes in the patterns in question, I thought the weight of probability favored the theory that the blue eyes were part of at least some of these patterns and not always a separate mutation. At the time I wrote:

The common theory in horses is that these W-series horses must have a splash mutation as well. And they may. Certainly there are far more mutations for white patterning than previous expected. I have long thought that the numbers of blue eyes on the dominant whites, particularly among the founder horses (ie., the horse that carried the initial mutation) were just too high for them all to happen to have a splash mutation as well. I did not have an exact number, though – just a sense that it was high. But the new sort gave me a number – six of twenty.

The argument that KIT mutations are “incapable” of producing blue eyes is based on information about how pigment is formed in the eyes of mice. I found the theory difficult to evaluate because the passages referenced by its supporters do not deal directly with eye color in horses, and I simply do not have the deep level of understanding of eye structure necessary to extrapolate beyond the specific subject, which was not the absence of blue eyes but the presence of dark eyes in a particular mouse color.

But perhaps more importantly, I just wasn’t sure that mice and horses were the same in this regard. In fact, I just wasn’t sure that eye color in various mammals might not be different in significant ways, just as other aspects of coloration vary between species. In that same post, I included a picture of what was then a newly-identified KIT mutation in dogs – the panda pattern in German Shepherds. I found the blue eyes on the founding dog particularly compelling because in dogs blue eyes are not generally associated with the common forms of white patterning, despite the fact that the mutations for most of those patterns have been found on MITF, which those of us more familiar with horses think of as the “splashed white” gene.

It seemed to me that these blue-eyed (KIT) German Shepherds, and the more common dark-eyed “irish marked” (MITF) dogs were a pretty good argument that when it came to eye color, there was probably some variation from the mouse model.

BlueNAmber4
The same pony, with one predominantly blue and one
predominantly gold eye (and an occluding spot over his blaze)

Some time after that post was made, I ran across a paper on Blue-Eyed White (BEW) alpacas. In that paper, the blue-eyed white phenotype was linked to the presence of two mutations to KIT. The alpacas, which were uniformly blue-eyed, were compound heterozygotes for two different white patterns (bew1 and bew2), both located at KIT. Since that time I have been able to confirm with one of the researchers that there were no mutations at the sites associated with splashed white in horses that could be correlated with the blue eyes on these alpacas.

Then just this month a paper was published linking the blue-eyed white phenotype in cats – called Dominant White (W) – to a mutation on KIT. In that study, the authors were quite clear about the connection between KIT and blue eyes.

In the population sample, we were also able to examine the correlation between genotype at the W locus and iris color. An individual that is homozygous W is much more likely to have blue iris, exhibiting odds 77.25 times larger than the odds of having blue irises of a genotype other than W/W (p < 0.0001).  An individual that is heterozygous (W/w+) also demonstrates increased odds of having blue iris (OR=4.667), four times larger than the odds of having blue irises of a genotype other than W/w+ (p=0.046) The odds of having blue irises of a wild type individual is 0.

With those two studies seeming to cast considerable doubt on the “Never From KIT” theory, I decided to contact one of the corresponding authors with some questions in hopes of getting a better understanding of this topic and of eye color in general. What I was told was that eye color is most likely a polygenic trait, and that it really does depend on the species, as well as the specifics of each particular mutation to KIT. That could explain why some KIT mutations are more prone blue eyes than others, as well as why there appears to be a higher incidence of blue eyes in homozygotes of some patterns.

BlueNAmber2
On this eye, the blue and gold portions are interspersed in such a way that the colors appear softly blended

BlueNAmber3
On this eye, there are fewer flecks of blue, as well as irregular patches of dark brown

So why does it matter if KIT mutations can produce blue eyes alone, or if they need a splashed white mutation? What purpose does this kind of information serve? The fact is that knowing the cause can help breeders more reliably get the outcome they desire, whether they wish to breed for or select against blue eyes. Likewise, breeders seeking to produce – or avoid – the splashed white phenotype need to know if blue eyes are always significant. If it is possible for some of the other patterns to produce blue eyes independent of a splashed white patterns, then assembling a herd of blue-eyed tobianos in hopes of developing a line of splashed whites is going to ultimately prove frustrating.

It is also true that quite a few blue-eyed horses have come back negative for the known splashed white patterns. It is likely that some of them have as-yet-unidentified splashed patterns. However, if some have blue eyes that are just a less common aspect of a pattern that is already identified, then knowing this could spare their owners time and money spent looking for something that is not really there. So while the subject is quite technical, it really does have a very practical aspect.

Continue Reading

“W” is for White-spotting

WhitePrince

The last few posts have focused a fair bit on the two KIT colors that, by long tradition, have their own names: tobiano and (dark-headed) roan. At the moment it is hard to imagine any problems with the terminology used for tobiano. The fact that it involves the same gene as sabino and dominant white has implications for breeders who are dealing with lines or breeds where those colors occur, but in most cases the fact that the horses are tobiano is still pretty obvious. The situation with true dark-headed roan is likewise pretty clear, at least visually. (The more confusing situation with roan genetics is a post for another day.)

The problematic terminology, at least at the moment, centers around the sabinos and the dominant whites. It appears likely that most, if not all, future discoveries are going to be given consecutive numbers in the W-series. Because that is true, it is helpful to keep in mind that the W in the name means “white-spotting” and not “white”. Related to that, the tendency to refer to all the patterns in the series as “dominant white” is probably not helpful, because while KIT mutations are expected to be dominant (that seems to be consistent across a range of species), many are not dominantly white. In fact, some are not white at all. Like the older category of overo, the white-spotting category may not really tell breeders what they need to know about a color; it is not specific enough.

From a practical standpoint, there are really three different categories that these white-spotted horses fall into. The first are the truly white – or at least nearly-white – horses. The horse above, White Prince, is an example of this type of pattern. In fact, his particular mutation (W2) is one that produces all-white horses consistently. There are others that produce a bit of color (usually on the ears or topline, often in the form of dark ticks or spots), but the overall impression is still that the horse is pretty white. A good example of a family with this near-white expression is that of the Arabian, R Khasper (W3).

The second category are the obvious pintos. These are the horses that are clearly broken-colored, with extensive white on the body. The average horseman, even one pretty familiar with color, would not call these horses white. Instead, they are most often called sabinos. The Thoroughbred family of Puchilingui (W5) and the Arabian families of Rhocky Rhoad (W15) and Fantasia Vu (W19) fall into this category.

SatoSide
Sato, Thoroughbred stallion with the W5 mutation

And finally there are the horses that would be thought of as having white markings. Not only are horses like this not obviously white, but many would not recognize this as a pinto pattern. The recently-discovered W20 mutation falls into this category.

MonaLisaGF
Mona-Lisa GF, German Warmblood mare heterozygous for W20

I say these are practical distinctions, rather than genetic distinctions, because most often breeders have a preference for one of these three phenotypes. Breeders of white horses, for instance, have struggled with the tendency their lines have for producing what they thought of as pintos. Meanwhile many breeders of pintos often want to avoid a horse like White Prince, which many colored horse registries view as “solid”. And patterns that produce flashy white markings, but that do not consistently produce white on the body, can be problematic in both solid and pinto breeds. That is why it is not necessarily helpful to categorize all the W mutations as “white”, nor for that matter to call anything that puts white on the horse “white patterning” – even if it is genetically accurate to do so.

The challenge, however, if that while those really are the categories that interest breeders, the genes themselves do not all fit neatly into just one of them. The relationship between the three basic phenotypes can only be called complex. In fact, there are intermediate colors that make the borders between the categories pretty fuzzy. Patterns like this one, for instance, sit somewhere between the white (and near-white) horses and the sabino patterns. For most horsemen, this is still very much a pinto pattern even though from a visual standpoint it is quite different from the pattern Sato has. Some of the draft and pony breeds, either by informal tradition or by actual rules, penalize horses with white on the body, but this kind of pattern might fly under the radar as “roan”. (If you nudge the contrast down a bit, that is even more likely to be true, and many individuals have less contrast than this particular horse.) Paint Horse breeders, meanwhile, would not only consider this a pinto pattern, but they would think of it as a fairly loud and obvious one. There are different traditions (and incentives) that play into how these pattern varieties are perceived.

ApacheLineSabino
A Thoroughbred from the Airdrie Apache line, which is widely rumored to have a mutation that has (to date at least) evaded detection

Here again are the images that given an idea of the range that white-spotting (W) can take: white, sabino-roan, sabino, and white markings. As pointed out, it can be difficult to draw a hard line between these four categories because one tends to blend into the next. But what makes the situation difficult to simplify is that there are different combinations that can produce the same color. Horses that look the same can be the product of a different genetic “recipe”.

KITrange
(Click to enlarge this graphic)

To illustrate this, let’s take a hypothetical all-white horse of unknown parentage. There are a number of ways, within the known mutations, to get that color. However, if she was then bred to an unmarked, solid stallion, the expected outcomes would be very different depending on why she was white. Here are some of the “recipe” options for her color and what they might mean for the resulting foals:

  1. She could be homozygous for Sabino1. In that case, she would produce 100% sabinos but no whites. Most would probably run towards the roany end of sabino expression.
  2. She could be heterozygous for a Dominant White variant that was inclined to be more truly white. She would then produce 50% white or near-white, and 50% solid.
  3. She could be heterozygous for a Dominant White variant that tended to “leak” color a bit more. She would be expected to produce 50% white, near-white, and sabino (probably more to the sabino roan end), and 50% solid
  4. She could be a compound heterozygote for two white-spotting genes. She would be expected to produce one of those patterns 50% of the time, and the other 50% of the time, but no whites like herself. If the other white-spotting pattern was something like W20, half the offspring might have white markings, but might not be readily identifiable as pinto-patterned.
  5. She might have a combination of patterns unrelated to the KIT gene which, when combined, give an all-white foal. In this scenario, she might produce quite a range of patterns either singly or in combination.

That means that without knowing what a white horse carries, you might have a horse that never produced its color, or sometimes did. Likewise it might never produce a truly solid horse, or it might sometimes do so. In the case of the fifth option, there could be entirely unexpected outcomes. It is not surprising then that white breeding programs established in Europe during the Baroque era typically ended in failure! Without a way to tell visually identical colors apart, the results would have seemed very unpredictable.

For many, it probably still seems this way. The fact is that what is true for one particular white-spotting mutation might not be true for the others. Each mutation requires its own explanation. With twenty different named white-spotting mutations, and many many more believed to exist, it is not surprising that many find that prospect discouraging. Because this is an area of ongoing research, it seems the most pragmatic approach is to not expect a perfect system for categorizing and naming these colors – at least not yet, when the picture is still incomplete.

In the next post, I want to focus on the effort to fund a study of the newest of the white-spotting mutations, W20. That study has the potential for expanding our understanding of these types of patterns, so I want to give it the attention it deserves.

Continue Reading