The science of horse genetics is based on an understanding of Mendelian inheritance
Gregor Mendel unravelled the laws of heredity - before either chromosomes or genes were known about - by observing the inheritance patterns of various characters of garden pea plants. Although Mendel established the principles of heredity through his work on pea plants they apply equally to horses. The science of horse genetics is based on an understanding of Mendelian inheritance.
Mendel’s successors expressed his findings in the terms of two laws of inheritance, usually using modern terminology that Mendel himself would’ve been unaware of. The observations now described make up the basis of what is often called Mendel’s First Law, or ‘The Law of Segregation’.
Mendel realised that inheritance could only be explained if the elements that determine characters (now known to alleles) existed in pairs. These pairs are separated in the gametes (sex cells), so that eggs and sperms carry just one member of each pair – whether any particular gamete carries one allele or the alternative one is determined by chance so that either is equally likely. When an egg and sperm fuse during fertilisation new pairs are formed. The next generation thus carries the new pairs brought together from the parents of this generation. Watch the fun video and look at the horse reproduction page to learn more about this process.
Applying Mendel’s First Law to horse genetics we can work out the possible characteristic of a foal from a particular mating, as long as one main gene controls that feature and we know the genotype of the parents at that locus. For example we’ll consider the probability of a gray foal from a gray mare and a chestnut stallion. The chestnut stallion is of genotype G+G+ (i.e. is homozygous for “non gray” alleles). The gray mare had only one gray parent herself and so must be heterozygous for the gray allele, i.e. of genotype G+GG. We can construct something called a Punnett Square to visualise this cross (always reminds me of strawberries!):
Altogether there’s a 75% chance of a gray foal (25%+25%+25%). Another way of saying this is that there’s a 3:1 ratio of gray to not gray. This ratio is what you normally expect when mating together two heterozygous horses (for any one character). Geneticists call this kind of mating a monohybrid cross.
Mendel wanted to know what happens when two pairs of contrasting characters are combined together in the same hybrid. We can illustrate what he found by considering horse genetics examples.
Let’s consider an example of the inheritance of two characters in horses. Say there was a breeder who specialised in producing chestnut paints. She has mares that are homozygous for the tobiano spotting allele (ToT). Since they are chestnuts they are also homozygous for the chestnut allele at the extension locus (e).
A neighbour has started specialising in champagne and now has a couple of stallions homozygous for the champagne allele (CHC). They are both golden champagnes and so are also true-breeding for the chestnut allele at the extension locus (e). (To read more about these genes read the sections on coat color genetics).
The breeder of paints took all her mares to the neighbours stallions. All the foals would have to inherit two copies of the chestnut allele at the extension locus since both mares and stallions were homozygous for this. The two breeders were interested in the inheritance of the spotting and champagne alleles in the foals.
The gametes from the spotted mares will be ToT CH+. The gametes from the champagne stallions will be To+ CHC. Therefore all the foals will be heterozygous at both loci, that is they will be (golden) champagne with tobiano spotting of genotype ToTTo+ CHCCH+.
The breeder kept two champagne tobiano fillies and an unrelated champagne tobiano colt foal. After a few years she intends to breed them together. By constructing a Punnett Square she is able to know the chances of producing champagne and tobiano champagne foals from these matings:
There is now a 9:3:3:1 ratio of champagne tobiano to chestnut tobiano to solid golden champagne to solid chestnut. This is the usual ratio expected when mating together two horses heterozygous for two particular characters. Geneticists call this kind of mating a dihybrid cross.
Thus the chance of a foal exhibiting both dominant characteristics is 9/16 (56.25%). The chance of a foal exhibiting one particular dominant characteristic and the other recessive characteristic is 3/16 (18.75%), with a 3/16 chance of exhibiting the alternative dominant and recessive characteristics. The chance of a foal exhibiting both recessive characteristics is only 1/16 (6.25%).
This is good news for our breeder who gets a premium for champagne horses (due to their rarity value) and for paint horses (due to the popularity of her well marked horses that also have good conformations and temperaments).
Taking each feature separately there’s a 3:1 ratio of champagne to non champagne (here chestnut) and a 3:1 ratio of tobiano to solid colored (or non tobiano). (Note that 12:4 = 3:1) As noted previously this is what you expect from mating together two horses heterozygous for any particular character (a monohybrid cross). For those of you with some mathematical ability you will notice that the dihiybrid ratio can easily be derived from multiplying out the monohybrid ratios:
(3+1) (3+1) = 9 + 3 + 3 +1
horse genetics and modified ratios
Mono- and di-hybrid crosses do not always give the classical 3 : 1 and 9 : 3 : 3 : 1 Mendelian ratios. This may be because the genes are linked on the same chromosome (for a dihybrid cross) or because they’re on the sex chromosomes (sex linked). In addition the expression of the genes in the phenotype can affect the ratios, especially where there is an interaction between genes.
In the examples so far only simple gene action was considered. For each gene one of the alleles has always been completely dominant over the other in the heterozygote. In the case of two genes the genes control two different characters and act separately from one another. Often the situation is more complex than in these examples.
In cases of complete dominance the heterozygote has the same phenotype as the dominant homozygote. This relationship between a pair of alleles provides the simplest situation for study.
With partial dominance (also called semi-dominance or incomplete dominance) the heterozygote exhibits a phenotype which is intermediate between the homozygous forms. There’s a well known horse genetics example, that of the cream dilution gene (the C locus).
Alleles at the C locus are responsible for the palomino, buckskin, smokey black, cremello, perlino and smokey cream. The two known alleles are designated C+ and CCr. CCr shows partial dominance and dilutes red to yellow in a single dose and to pale cream in a double dose.
Horses with a chestnut base color and genotype C+C+ are chestnut, while those of genotype CCrCCr are cremello. Horses of genotype C+CCr are palomino, a color intermediate between the phenotypes for the homozygous forms. Similarly horses with a brown or bay base color and genotype C+C+ are brown or bay, while those of genotype CCrCCr are perlino. Horses of genotype C+CCr are an intermediate color of buckskin. Cream dilution can have a very subtle effect on black pigment and horses with a black base color are diluted to smokey black (C+CCr) or smokey cream (CCrCCr). The wild-type C+ allele is effectively recessive since it needs to be homozygous for there to be no dilution of the base color.
The phenotypic ratio for a monohybrid cross, e.g. a cross between two palominos, is modified to 1: 2: 1 chestnut: palomino: cremello. This is the same as the genotypic ratio since the heterozygotes are of a separate phenotype (i.e. heterozygotes are palomino and not cremello, as they would be in the case of complete dominance of the CCr allele).
In co-dominance both alleles are expressed in the phenotype and the heterozygote has the characters of both parents. The inheritance patterns for co-dominance are similar to those for partial dominance.
The best known examples of co-dominance are those of the blood groups, for example the AB blood group where animals with blood group AB have both A and B type antibodies in their blood. These aren't specifically horse genetics examples but many mammals have blood groups in common and their genetics is same.
There are also loci with alleles made up of non-coding DNA, i.e. DNA that isn’t responsible for any external phenotype characters. Such loci can have lots of different alleles (because they don’t “do” anything mutations aren’t harmful to them and lots of variation can build up without natural selection throwing it away). Some of these loci are useful as genetic markers in molecular genetic studies. They have, for example, been used to find genes for some important horse genetic disorders, such as equine combined immune deficiency disorder (equine CID, discussed elsewhere in relation to linkage). These loci have phenotypes only at the level of a molecular genetics test, but there inheritance is nevertheless co-dominant in as much as both alleles at a locus can be detected.
There are a few well known horse genetics examples of lethal alleles, including the white allele, the overo allele and possibly some variants of the roan allele. Lethal alleles result in modified ratios among surviving foals.
Two alleles are known for the gene for white coat colour, symbolised WW and W+. Most horses are not white and have genotype W+W+. The WW allele is rare in most breeds of horse, but occurs in Tennessee Walking Horses, American Albinos and Miniatures, and rarely in Arabians, Standardbreds and Thoroughbreds. Horses with the WW allele are dark-eyed horses with white coats. WW is dominant over W+, so that horses of genotype W+WW are white.
No horses are known with the genotype WWWW. Breeding between white horses always produces some coloured foals, indicating that the horses are heterozygous. It would seem that embryos or foetuses homozygous for allele WW die early in gestation and are then either resorbed or miscarried. WW is therefore acting as a recessive lethal allele. We therefore say that the allele WW is dominant visible and recessive lethal.
The following diagram shows how the standard 3:1 ratio of a monohybrid cross between white horses is modified to a 2:1 ratio typical of recessive lethal genes.
A second well known horse genetics example of a lethal gene is that which causes the white pattern in overo horses. There are various different genes that cause white coat patterning in paints or coloured horses, and overo is genetically distinct from other white patterns such as tobiano and sabino. Overos are heterozygous for a gene that is lethal when homozygous. Thus the overo allele (OO) is dominant for colour pattern but has a recessive lethal effect. The allele shows pleiotropy, which means they have more than one effect on the phenotype (affect more than one character). Homozygous foals (OOOO) are all-white with blue eyes and die of complications from intestinal tract abnormalities. Both melanocytes (pigment cells) and ganglia (nerve cells) are migratory cells that originate from the same area of the developing foetus known as the neural crest. The all-white foals lack both pigmentation and nerve cells in the intestinal tract (aganglionosis).
When two heterozygotes are bred together an average of 25% of foals are lethal white. Surviving offspring are either overo or solid coloured. Matings between solid and overo horses result in solid and overo foals in approximately equal numbers, with no lethal white foals. Occasionally, however, horses without noticeable body spotting patterns have sired or produced lethal white foals. The overo spotting pattern is phenotypically heterogeneous (i.e. it varies a lot) and it is possible that such horses show insufficient white spotting for registration purposes, even though they have the overo genotype. (Another possibility is that the overo mutation occurs “de novo” in the gametes of one parent.) Breeders would obviously like to be able to recognize horses at risk of producing lethal white foals. Until recently, there has been no reliable way to identify which horses have the gene associated with lethal white overo (LWO). Now, however, there is a molecular horse genetics test.
It used to be thought that the roan allele was an example of a dominant visible but recessive lethal allele. Breeding roans together was not advised and it was thought that embryos homozygous for roan were reabsorbed in very early pregnancy. This has now been shown not to be the case and some homozygous roan stallions have recently been identified using horse genetics molecular technology.
Epistasis is a gene interaction where an allele or alleles at one gene masks the phenotypic expression of alleles at a second gene. The phenotype is governed by the masking gene when both are present together in the genotype. A genotype that masks another’s expression is said to be epistatic, while the gene whose expression is masked is said to be hypostatic. There are several examples of epistatic relationships in horse genetics, especially between genes that govern coat color.
Epistatic alleles may be recessive or dominant. If they are recessive then individuals homozygous for the epistatic allele are of the same phenotype regardless of the genotype at the second gene. Alternatively epistasis may result from the presence of a dominant allele, which also conceals the genotype at the second locus.
Recessive epistasis is shown by the extension and agouti genes, which account for the differences between black, bay, brown and chestnut horses. The recessive alleles of the extension gene are epistatic, alleles of the agouti gene are hypostatic. The alleles of the extension gene extend (E+) or diminish (e and ea) the amount of the black eumelanin pigment in the coat, with opposite effect on the amount of the red pigment phaeomelanin. The dominant alleles of the agouti locus cause the distribution of black hairs to be restricted to the points (e.g. lower legs, mane, tail and ear rims). Horses homozygous for the recessive allele Aa are uniformly black. Since there are no black hairs in chestnut horses the agouti gene can have no effect on the distribution of black in these horses. Thus the genotypes ee, eaea and eea conceal the genotype at the agouti locus: whatever the genotype at the agouti locus these horses are always chestnut (or sorrel), or some colour derived from chestnut, such as palomino or red roan.
We can demonstrate the affect of epistasis on genetic ratios through a horse genetics example. Say we had a bay mare and stallion who were both of genotype E+e at the extension locus (E+ causes the production of the black eumelanin pigment) and of genotype AAAa at the agouti locus (which controls the distribution of black pigment).
The gametes may now be of four types, any of which are equally likely: E+AA, E+Aa, eAA or eAa. The possible outcomes of the cross can be seen from a Punnett square:
Instead of a 9:3:3:1 ratio there is a 9:3:4 of bay: black : chestnut. The agouti allele in the chestnut horses is irrelevant to the phenotype since there is no black pigment to distribute, either uniformly or in the points.
A horse genetics example of dominant epistasis is that of the gray allele. Horses with at least one copy of the allele GG go through the graying process regardless of the genotype at the other genes controlling coat color. They may be born chestnut, bay, buckskin or any other color but they will steadily turn gray over time, and eventually may turn almost white.
Because of the gray allele being dominant heterozygous gray horses can have foals of other colors, depending on their genotype for the other color genes. Consider, for example, that we had a gray heterozygous mare and stallion (both of genotype G+GG) who were also heterozygous for the allele e, that causes chestnut when homozygous (i.e. they are therefore both of genotype E+e). The gametes from each horse would now be of four types, any of which are equally likely: G+E+, G+e, GGE+ or GGe.
The Mendelian segregation of two or more pairs of characters occurs because the genes controlling the characters are located in different chromosomes pairs. The alleles of one gene separate independently of the alleles of the other gene (or genes) so that any one particular allele is as likely to be recombined with either of the possible alternative alleles at the other locus (or loci).
If all genes were on separate chromosomes then independent segregation would provide us with an adequate description of heredity. We now know, however, that chromosomes are made up of linear sequences of large numbers of genes all linked together. Linked genes don’t behave independently of one another in their inheritance. To understand horse genetics well we therefore have to consider how such linked genes are inherited.
Linkage is the association of genes in their inheritance due to them being located on the same chromosome. The same term may also be used to describe the characters determined by the linked genes (the characters may be said to show linkage). When we look at the inheritance of characters determined by linked genes we do not see the Mendelian ratios typical for two genes inherited independently of one another.
I will describe linkage through the horse genetics example of the gene for equine combined immune deficiency disorder (equine CID). The recessive allele of this gene causes equine CID and its characterisation was one of the first success stories of equine molecular genetics. It’s thought that about ¼ of all Arabs carry the gene for equine CID (McGuire and Poppie, 1973), presumably due to inbreeding within the breed. Equine CID results in a deficiency of the immune system and foals born with the condition usually die within 3 months of birth.
The process of finding and characterising the gene involved identifying “marker” genes linked to the equine CID gene. (Looking for an unknown gene is a bit like looking for a needle in a haystack when you’re not quite sure what the needle looks like but you know it’s attached to some cotton – you find the cotton first! Briefly the idea is to start a search by first finding out what the gene is close to that you already know about, i.e. which known genes it’s linked to).
If 2 loci (the CID gene and a marker) are closely linked only 2 types of gametes would be expected. This was seen to be the case for the genetic marker called HTG8.
Suppose the equine CID gene is denoted D and the marker M, with d being the disease allele and m the other marker allele. If M and D were close on one chromosome and d and m on the other we would expect gametes with genotypes DM and dm in roughly equal proportions (i.e. only of the parental genotypes). If a carrier stallion was mated with a non carrier mares we’d see the following genetic ratios:
If a carrier stallion was mated with a carrier mare we’d see the following genetic ratios:
This ratio is like that which you’d expect from a monohybrid cross since the linked genes are inherited as a single unit (carriers are phenotypically indistinguishable from normal foals unless a molecular genetics test is done to identify the equine CID allele or breeding reveals carrier status).
If the two genes were independently assorting they would not be genetically linked at all. Then 4 types of gametes would be expected in equal proportions. The d allele would then be as likely to segregate with either the M or the m allele during gamete formation, the same is true for the D allele.
Many genes are partially linked. In such cases we get 4 gamete genotypes, but not in equal proportions (i.e. the genotype and phenotype ratios do not conform to those for dihybrid crosses). This occurs because the maternal and paternal chromosomes come together and swap some genes in a process called crossing over. This gives rise to new combinations of characters and so is a source of genetic variation. The farther apart two genes are on a chromosome the more likely there is to be crossing over between them (since only one or a few crossing events occur per pair of chromosomes).
The actual proportions of gamete genotypes depends on the distance between the two genes on the chromosome. The following represents what might happen during gamete formation. Each chromosome is made up of two identical strands, each called a chromatid. The chromatids separate from one another during gamete formation and later replicate to form new chromosomes:
If there is always crossing over between two genes (because they’re far apart on the chromosome) the genes aren’t considered to be genetically linked and the ratios expected for such a di-hybrid cross are the same as those for independent segregation.
The genotype and phenotype ratios for partial linkage are therefore somewhere in between those for complete linkage (with only parental genotypes and no recombinant ones) and independent segregation (when recombinant and parental genotypes are equally likely).
Because of the differences in expected genotype proportions according to linkage relationships horse genetics scientists can identify marker genes that are very close to genes for genetic disorders, such as equine CID. This makes it easier to find, isolate and characterise those genes using molecular horse genetics procedures. This can lead to horse genetics tests, and to improvements in the understanding and treatment of genetic disorders.
The extension coat color gene (E locus) is linked to the genes for roan (RN) and tobiano (To). The dominant extension allele is E+ and horses with this allele produce black eumelanin pigment, resulting in the horses of black, brown or bay and their derivatives, depending on other genes. Allele e is recessive and horses of genotype ee are chestnut (including sorrel) or some derivative of that color (such as palomino). The dominant allele at the tobiano locus is ToT, with horses of genotype To+To+ being solid colored rather than white spotted.
For a particular horse genetics example we will consider what happens in di-hybrid crosses between two black tobiano horses (i.e. between horses of genotype Ee ToTTo+). Because of crossing over any particular animal may have chromosomes with the genotypes E ToT, E To+, e ToT or e To+ – and this will affect the expected ratios of the different phenotypes and genotypes. We will assume that both our horses have chromosomes of genotype E ToT and e To+. We will further assume that both horses are of genotype AaAa at the agouti locus (see section on color genetics or black horses for more information). Lets say for arguments sake, that 20% of gametes had recombinant genotypes. The following shows what we expect from such a mating:
Black tobiano: 16%x3 + 4%x4 + 1%x2 = 66% Black: 4%x2 + 1% = 9% Chestnut tobiano: 4%x2 + 1% = 9% Chestnut: 16% (Total 100%)
From this you could see that, on average, about 2/3 of the foals would be expected to be black tobiano. Put another way any particular foal has a 66% chance of being black tobiano. If there were fewer recombinant chromosome types (because the genes were closer together) then fewer black and chestnut tobiano foals would be produced.
The alleles may be linked differently in different horses, and particularly in different breeds of horses, so that, for example, ToT e and To+ E might be the parental types. This is the case in the example that follows. There are also assumed to be fewer recombinant chromosome types to show how this can affect ratios:
In this case you can see that about half the foals (on average) are expected to be black tobiano and half either balck or chestnut tobiano. Another way of saying this is to say that for any particular foal the chances of it being black tobiano are about 50%. Very few foals would be chestnut – or to put it another way the chances of getting a non-tobiano chestnut foal are small.
Sex linked genes are on the sex chromosomes, that is to say the X and Y chromosomes (non sex chromosomes are called autosomes). In actual fact most sex linked genes are on X chromosomes. The Y chromosomes are smaller than X chromosomes and contain very few genes.
Since the X and Y chromosomes are largely non-homologous males have only one allele for almost all of the sex-linked traits. They are said to be hemizygous for these traits. This results in recessive alleles being expressed without them needing to be homozygous (which they would have to be in females). Especially when alleles are rare, such as those for genetic disorders, they tend to be in the heterozygous form in females (since the other allele is statistically more likely to be a common allele than a rare one – the same is true for autosomal traits). This leads to recessive sex-linked traits being more common in males than females, for example, some kinds of color blindness. It also leads to altered genetic ratios compared with those for autosomal traits.
The inheritance of the sex chromosomes is shown below, and we can see that – on average – we expect a 50:50 ratio of males to females, even though in nature (and in domestication) a few stallions tend to have hareems of mares.
An example of a sex-linked gene is that causing haemophilia A (factor VIII deficiency). Although the example isn’t unique to horse genetics, it has been reported rarely in Thoroughbreds, Quarter Horses and Standardbreds (Archer, 1961, Henninger, 1988, Hutchins et al, 1967).
References and further reading for the horse genetics section
There are many horse genetics research papers in scientific journals, but finding good up-to-date horse genetics books that don't cost the Earth is more difficult. The principles of genetics apply generally to all organisms, and so apply to horse genetics. A good general textbook can therefore be helpful to those particularly interested horse genetics.
Archer, R.K. 1961. True haemophilia (haemophilia A) in a Thoroughbred foal. Veterinary Record 73, 338-340.
Bailey, E., Reid, R., Skow, L.C., Mathiason, K., Lear, T.L. and McGuire, T.C. 1997. Linkage of the gene for equine combined immunodeficiency disease to microsatellite markers HTG8 and HTG4; Synteny and FISH mapping to ECA9. Animal Genetics 28 (4), 268-273.
The international Equine research site at http://www.projects.ex.ac.uk/equinet/abstracts.html. Has some interesting horse genetics information. Last accessed on 8/10/2005.
A useful horse genetic book
Gower, Jeanette. 2000. Horse Colour Explained. A breeders perspective. The Crowood Press. ISBN 1 86126 384 8. My personal favourite horse genetics book, although it only deals with color and pattern. The photos are superb.
Bowling, Ann T. 1998. Horse Genetics. CAB International. ISBN 0 85199 101 7.
This page might seem long, especially if you're new to horse genetics. However horse genetics is a big subject! New horse genetics research findings emerge often, including on genetic disorders, mapping the horse genome, horse genetics and evolution and conservation horse genetics. Over time I intend to extend this horse genetics site to cover more varied aspects of horse genetics, including some aspects of molecular horse genetics and conservation horse genetics.