Brown out!

The most common emails I receive are about hair and eye color, and of these the most frequent source seems to be from individuals in interracial relationships. Quite often they are curious as to the possible outcome of their offspring's phenotype. Sometimes they wonder why their offspring looks the way he or she does. On one disturbing occasion someone was appealing to me to clear up the suspicion of non-paternity because of the unexpected outcome of the offspring's appearance! Today I received this email:

I have a 19 month old son who is very light skinned, blond hair and blue eyes. My husband (American) is bi-racial. His mother is white and his father is black.

Myself (German), I'm white, same with both of my parents and grandparents. I'm a redhead with greenish brown eyes, my husband has black hair and brown eyes.

What are the odds for this to happen, for us to have a white baby with blond hair and blue eyes?

This question is very common: why does my baby look so white? Succinctly the child's ancestors are mostly white (American blacks are on average 20% white, so it is likely that the child is more than 75% European in terms of recent ancestors). I've offered more detailed expositions on black white twin pairs. The underlying logic is that genetics is discrete, not blending. This is the the insight and power which Mendelian principles introduced to our understanding of evolutionary process. All that being said the above posts and these sorts of jargon-laden assertions really don't mean much to many people. So I'm going to answer the email above in some more detail below.

But first, one particular issue that I want to bring up. Some of the emails I receive imply strongly that people are surprised at the relatively lack of potency of colored blood, especially black blood. I think what's being assumed as a background condition is the rule of hypodescent, which assigns a mixed-race child to the identity of the lower status parent. One drop of black blood was enough to render someone a black person in the Untied States. It seems to me that many people are somewhat shocked that an individual who many would say is black, as the man above, could be the father of white offspring. After all, is black blood not strong? Well, I hate to disappoint Afrocentrists but there's nothing special about black blood. Add enough creme and it will give way. The problem is that quite often mixed-race individuals are identified as if they are the socially subordinate race, so the logic would imply that their own child would inherit that subordinate race. But when the child "looks white," there is some cognitive dissonance and we have to unpack our assumptions.

This is not to say that the assumption of hypodescent is necessarily racist. Quite liberal people now accept it because it is a social convention to allow someone who is of mixed-heritage to identify as only of the subordinate race. For example, most people accept Halle Berry as a black woman even though she has a white mother. Now, if Berry asserted that she was a white woman because she had a white mother I assume most people would look at her a bit strangely. As it happens Berry is now pregnant and her child will be genetically 3/4 white, and there is a strong likelihood that he or she will be able to "pass" if they so wish (if they insist on being identified as a black American I suppose most individuals will allow them to continue to do this). Consider the daughter of the actress Victoria Rowell, who is biracial. She looks white, and her ancestry is 3/4 white. In the past in the United States individuals such as Maya Fahey, Rowell's daughter, would have passed into the white population or been absorbed into the black population. But this is not necessarily a choice she needs to make today, and it is not a choice that is even as important as it was in the past. I doubt that with a brother who is 3/4 black Maya Fahey will wish to pass and disappear into the white population and disavow her African American ancestry.

In any case, enough sociology. What are the chances that two non-blue-eyed people will give rise to a blue-eyed child? First, let us remind ourselves that children tend to grow darker with age, and often blue-eyed babies turn brown-eyed at some point in their lives. But this child is nearly 2 years old so let's take development off the table. Eye color is controlled mostly from one gene, so let's assume that it is a single locus trait. The child has blue-eyes, but neither parents have blue-eyes. So I would say that the likelihood here is on the order of ~1/4, give or take, that future children will also have blue eyes. It's a mostly recessively expressed trait, so the ratio would emerge naturally since we know parents are not blue-eyed, so they are heterozygotes. The mother is a red-head, so it is likely she has loss of function on MC1R. The child's hair has a blondish cast. If the father's white parent was fair-haired then the combination with a red-haired women can naturally resulted in the emergence of a fair-haired offspring. The genetics is a little tricky here, I'm not comfortable with a single gene approximation. That being said, red-hair implies very little production of eumelanin, so any moderate skewed sampling toward the loss of function alleles which the father likely carries from his white parent could result in a fair-haired child. If I was a betting man I'd give the odds at around 1/3 or so, with most of the variation in hair color coming from the sampling of the father's genes. Finally, in regards to skin color the red-hair and fair skin of the mother means that this is likely again decided by sampling of the father's genes. If he is mixed-race he likely carries functional and non-functional copies of SLC24A5, SLC45A2 and TYR in heterozygote genotypes. There's a 1/8 chance that on these three loci only European variants will be passed along and a 1/8 chance that only African ones will be passed along. There is another gene or two of some significance, and it is very likely that the child has a non-functional MC1R from the mother already. The child doesn't seem as fair as the mother, so let's assume that it received one functional copy from the three genes listed above, that gives 3/8 shot for that outcome. There's a 1/2 shot that the child could be darker on these genes (a 1/8 chance it could be lighter, receiving no functional copies).

I throught out some general probabilities based on inference of genotype from phenotype, what I know about family history, and the sketchy outlines of the genomic architecture which we're starting to understand now. For example, normal human variation for skin color is mostly controlled by 4-5 genes, so I generally just pretend as if it's a binomial distribution. Eye color is 3/4 controlled by one locus, so I pretend it's single locus, though that's mostly useful with blue vs. non-blue distinctions. Recent work on hair color has added a lot more detail which I should have cranked through, but I think the rough estimates would work. Of course, in the near future the parents would just get sequenced on the appropriate loci and they could get a phenotypic expectation and variance.

In short, I don't think that the baby is that unlikely.

Note: The correspondent encouraged me to share their story and post the photo. Just so you know!