Westeros is clearly not the place to nitpick about scientific realism: the very turning of the seasons seems to answer to nothing so much as the whims of gods and monsters. But I can’t really help myself. There are two wonderful things about genetics in Westeros: first, compared to many other fictional universes, we have a lot of intricate family trees upon which to base genetic observations. Second, sometimes it really works, and sometimes it really doesn’t.
I’m going to start this with something that is probably very boring but can be a great segue for something very interesting: hair color and the Baratheons. The short story is that every Baratheon ever has dark hair. The boring explanation: there aren’t that many generations at play, and there were even fewer direct descriptions of appearance, and since dark coloration tends to be dominant over light coloration, it’s mostly coincidental (expect a post later about whether Ned actually had enough information about Robert’s bastard children to tell whether Tommen, Myrcella, and Joffrey were trueborn). The interesting explanation: no, seriously, every Baratheon ever – as in, anyone who has a single Baratheon ancestor, has dark hair.
When you think about genetics of relatively simple traits like hair color, everything is controlled by a few genes and inheritance is roughly Mendelian: considering the oversimplification of brown (B) and blond (b) hair, if you kept injecting blond (bb) parents you’d eventually get some blond (bb) offspring. Anyone with brown hair and one blond parent will be Bb; if they have children with another blond you’d expect half of those children to be blond. So about half the time, a child with a blond mother and one blond paternal grandparent (or a blond father and one blond maternal grandparent) will be blond.
Basically, after enough generations of marrying into families like Lannister and Targaryen, we would expect that some of the Baratheon kids would have blond hair. But they don’t. That doesn’t sound like Mendelian inheritance. What it does sound like is another phenomenon, called paramutation. Paramutations are interactions between chromosomes that lead to heritable changes. And yes, that’s a direct violation of Mendel’s laws: traits are supposed to segregate and separate, and any one trait isn’t supposed to have any effect on the heritability of any other. But it happens. Wacky, right?
It’s hard to understand what I mean by “interactions between chromosomes that lead to heritable changes” without an example. And most of the examples we know of right now are in plants. So let me take you through one of the best examples of paramutation, p1 in corn, and I’ll compare it to Mendelian inheritance along the way.
p1 is a gene that controls kernel color. P1-rr is an allele of p1 that yields dark red kernels. P1-rr’ is an allele of p1 that yields mostly yellow kernels. When you cross P1-rr/P1-rr plants with dark red kernels with P1-rr’/P1-rr’ plants with mostly yellow kernels, you get ears with mostly yellow kernels. So far this is just like Mendel’s pea plant experiments: a Y/Y pea plant with yellow peas, when crossed with a g/g plant with green peas gives Y/g plants with yellow peas. Now let’s consider where things diverge: if you take that Y/g plant with yellow peas and cross it with another g/g plant with green peas, half of the time you get Y/g plants with yellow peas and half of the time you get g/g plants with green peas. But if you take the plants with yellow kernels from the P1-rr/P1-rr X P1-rr’/P1-rr’ cross, and cross them with red-kerneled P1-rr/P1-rr plants, you again get ears with mostly yellow kernels. And since each kernel is an individual, we can screen hundreds and thousands of individuals very quickly: instead of 50% of the time, we see yellow kernels 93% of the time.
Not weird enough for you yet? It gets better: P1-rr and P1-rr’ have exactly the same sequence. Under certain conditions, with certain viral sequences in a cell, P1-rr will even spontaneously turn into P1-rr’.
So if it’s not Mendelian, and it’s not DNA sequence, what is going on? The P1-rr sequence has repetitive, virus-like, sequences flanking p1. Most repetitive sequences like that get “turned off.” In the P1-rr form, it’s still turned on, though, so p1 still gets expressed and you still get red pigment. In P1-rr’, it’s off, and you don’t get any p1 or any red. The presence of the P1-rr’ allele in a cell will trigger the cell to turn off any genetically identical sequence — the P1-rr allele.
One great part about this story is that we’re still not exactly sure how this would happen, molecularly. Why does a cell know to turn off P1-rr? One of the simplest explanations is that instead of making the p1 gene, P1-rr’ transcribes short RNA molecules which can silence repetitive elements. Those RNA molecules don’t just recognize P1-rr’, but also P1-rr, and so silence P1-rr and turn it into P1-rr’.
Which brings me back to hair color. If we draw the analogy out farther, Baratheon hair color alleles would stably silence every other allele they came in contact with, resulting in a heritable suppression of any kind of hair pigmentation. The only problem with this pattern is it wouldn’t result in dark brown hair like Robert Baratheon’s. If “Baratheon” individuals lacked all (or most) hair pigment, they would likely have colorless – silvery white – hair, rather like one other powerful family in Westeros.