Debunking X Factor

The “X factor” is an influential theory about the inheritance of heart size in thoroughbred horses. It was originally put forward by Marianna Haun (pictured), a journalist who covered horse racing before becoming a breeding consultant. She came upon an old observation from racehorse breeding circles that some very good male horses failed to produce any quality offspring but that these same horses had successful grandoffspring through daughters. Such horses are known as broodmare sires, because while their offspring don’t have success on the track, their daughters have success in the barn, as broodmares. Heart size is an obvious predictor of racing ability and seems to be involved in this broodmare sire effect, at least from the anecdotal evidence that Haun has compiled. Secretariat had a massive heart, his offspring not so much, but his grandoffspring through daughters had massive hearts, too, so the anecdotes go. I’m no fan of anecdotal “evidence” but I find no harm in letting one’s hypotheses get a lead out in front of the real evidence. What I’d like to do now is present Haun’s hypothesis, the “X Factor,” along with an alternative that I think fits the anecdotal evidence better. First, I ought to explain that there are several different routes of inheritance for genes in a mammalian genome. The most common is for a gene to be autosomal. Your autosomes come in pairs and you get one each, symmetrically, from mother and father. You, as a human, have 22 pairs of autosomes (44 chromosomes) numbered 1 through 22. A horse has 31 pairs of autosomes (62 chromosomes), numbered 1 through 31. Another possible route is sex-linked. You, as a human, also have a 23rd pair of chromosomes, or sex chromosomes. If you are female, you have two X chromosomes, one each from mom and dad. If you are a male, you have one X (from your mom) and one Y (from your dad) [X&Y Chrom shown]. There is an asymmetry in the inheritance of sex chromosomes that makes them radically different from autosomes. Men should notice that they couldn’t get a Y from mom, because if she had a Y, she’d be Dad (the Y would make her a man). So, while the X in males always comes from mom, the Y always comes from dad. Taken further, this means that the Y chromosome from dad came from his dad, who got it from his dad, etc. On a pedigree, you can follow the top line to trace the Y chromosome back through its ancestry. (As an aside, in humans, you could potentially follow the same ancestral route for last names – as though the last name was a gene found on a Y chromosome! – though, in actuality, this inevitably fails because of name changes.) You can’t follow the bottom line of a pedigree to trace the ancestry of the X, however, and the pattern of inheritance of the X is much more complicated that that of the Y or even that of the average autosome. Most pedigree software will highlight the genealogy of the X chromosome for you. Horses also have sex chromosomes, an X and a Y, and ~5% of the genes in the horse genome are found on the X, an estimate that is fairly steady across all mammals (in Drosophila [pictured], it averages ~20% but can go much higher). Haun hypothesized that the broodmare sire effect is due to the inheritance of a single gene on the X chromosome. This gene exists in two varieties in the population: one that leads to the development of a small heart and another that leads to a heart the size of Secretariat’s, say. If such a gene is X-linked, this explains why it is more heritable from the maternal grandfather than from the paternal grandfather. A paternal grandfather, let’s call him Pat, sires the father, who receives the Y from Pat – not Pat’s X. A maternal grandfather, let’s call him Matt, sires the mother, who receives Matt’s X. Thus, for the sex chromosomes in the grandoffspring generation, the paternal grandfather can only contribute genes from his Y (and not his X) but a maternal grandfather may contribute genes from his X chromosome (but not his Y) through his daughters. So, the X factor hypothesis explains one feature of the broodmare sire effect: individuals will resemble their maternal grandfathers more than their paternal grandfathers for X-linked traits because only the maternal grandfathers can pass on X chromosomes to their grand-sons and grand-daughters. The aspect of the broodmare sire effect that is not explained by X linkage is the averageness of the daughters of broodmare sires. These daughters certainly inherit the hypothesized gene variant on the X from their father. This raises a question: Why don’t these daughters benefit from it? Why don’t they have large hearts and clean up at the track like their fathers? Haun’s reply would be that the variant of the gene that underlies large hearts is recessive and is masked in daughters by the presence of the other X chromosome, which daughters get from their mother. This explanation only works if the other, maternally inherited, copy in daughters is always the small heart version of the gene and if this small heart version is dominant. But if both of the X chromosomes that a daughter inherits have the recessive variant of the gene, then the large heart effect can’t be masked by a dominant version and these daughters should express the large heart trait and should be successful runners. We can assume that the large heart version is at a frequency that is appreciable enough so that we would obtain such “double-copy” females enough times to notice a measurable effect (Haun even admits of there being double-copy mares). The effects of genes on the X chromosomes, even if they are recessive, can’t miss the daughter generation entirely (Think horse's by Storm Cat out of a AP Indy mare). This is a glaring weakness of the X factor hypothesis. An alternative that I’d like to offer is capable of explaining the higher heritability on maternal rather than paternal grandsires and is also capable of explaining the averageness of daughters. This alternative hypothesis is that the gene involved in heart development is imprinted. Imprinting of genes is yet another route of inheritance that differs from autosomal and sex-linked. If a gene is imprinted then it is expressed in a parent-specific manner. Some imprinted genes are exclusively maternally expressed in the population while others are exclusively paternally expressed. To the best of our knowledge, there exist on the order dozens to hundreds of imprinted genes in the mammalian genome and virtually all of them are on the autosomes – not the X or Y chromosome. Let’s take a hypothetical example, the MEG55 gene. This is a maternally expressed imprinted gene. This means that the version of this gene that you are using to make the MEG55 protein is the version that you inherited from your mother. The version you inherit from your father is there in all of the cells of your body but it is kept silent because of a reversible chemical modification to the DNA of this particular gene. It’s reversible because, if you’re a female, when you pass on one of your two versions to your offspring, you will be passing on an active version. If you are a male, you will be passing on one of your versions in a silenced state to your offspring. These imprints don’t change the DNA sequence. They merely prevent the body from using that version of the gene for one generation. Consider the possibility that there are two varieties of Meg55 gene in the thoroughbred population: one that gives rise to large hearts and another that gives rise to small hearts. And let’s say that the version that Secretariat inherited from his mother was the large heart version. This would explain why Secretariat had such a large heart and was so successful on the track. But when Secretariat had offspring, all of them received their version of his Meg55 gene in a silenced state (because it is paternally derived in them). So, they didn’t resemble Secretariat at all for the heart trait, unless by chance. This feature of imprinting is capable of explaining the averageness of the offspring of broodmare sires. But, Secretariat’s daughters could pass an active version of the Meg55 gene to their offspring, and there was a chance that this version that Secretariat’s grandoffspring receive in the active state would be the exact copy that Secretariat had in the active state and used to power himself to all those victories. For imprinted genes it is impossible to resemble any ancestors of one side of the pedigree. For maternally expressed genes, like our hypothetical Meg55, there will be no resemblance of individuals to their father’s family. Likewise, for paternally expressed genes, there is no resemblance between individuals and their mother’s family. Another feature of imprinted genes is that their effects on traits of interest (like speed, heart size, etc.) are capable of skipping generations entirely. Thus, imprinting fits the anecdotal pattern for heart size better than X-linkage does because it is capable of explaining the two main features these anecdotes uncover. Can we discriminate between these two hypotheses for the inheritance of heart size? Indeed we can. Breeding geneticists can use the patterns we measure and test whether they fit an X-linked or an imprinted pattern of inheritance better. All that is needed is heart size information on a large number of related horses with accurate pedigrees. A recent paper by Alastair Wilson and Andrew Rambaut used such breeding genetic models to tease apart whether it’s the genes or the training that is more important to success in thoroughbreds (it’s the training). I’d encourage them to tease apart whether it’s X-linkage or imprinting that explains the pattern of heart size inheritance in thoroughbreds (or, of course, whether the old anecdote about an unusual pattern of inheritance for heart size should be laid to rest). The Doctor recently recieved his PhD from Harvard University in Biology