6
$\begingroup$

Many questions already exist regarding dominance/recessive relationships, see e.g. here. I am asking this question because I have often wished to have it to refer to, and also genuinely curious whether someone has a clear answer grounded in evidence.

A common misconception that one encounters in genetics (especially as received from a high-school level class) is that all genes affecting a phenotype follow the classic Mendelian relationship of a dominant / recessive inheritance pattern yielded by diploid organisms. See for example this recent question.

It follows that many consider this relationship to be foundational to genetics rather than largely the result of ascertainment bias. In contrast, most quantitative genetics models perform well assuming linearity / additivity, i.e. "co-dominance". It is not clear whether this is an outcome of the real state of affairs, or simply that the additive model is "good enough" for summarizing the complex reality of underlying epistasis, dominance, etc (I believe "good enough", but that's somewhat philosophical).

My question is this: what proportion of variants with quantifiably large phenotypic effects fit a dominance model of inheritance?

The question is muddied somewhat by incomplete (but not co-) dominance, or the possibility of multiallelic loci with many alleles with a complex dominance hierarchy, as in this example. But I'd be satisfied with a rough quantitative estimate.

My vague sense is that most variants are better explained by an additive model, if only because their effects are rather small. In other words, I think that our fixation on dominance is a result of ascertainment bias inherited from Mendel. We then teach it to early students because it's simple and illustrates some important features of genetics. Am I wrong?

$\endgroup$
2
  • 1
    $\begingroup$ This seems like it would be exceptionally difficult to quantify. By far the most common cases of apparent dominance are going to be those involving enyzmes where a recessive allele is somehow non-functional, yet the product isn't limiting or negative feedback is present such that haplosufficiency is observed. Even then, though under some circumstances you may see some evidence of haploinsufficiency, or there may be some other allele that shares a phenotype with the "true" dominant allele, yet shows haploinsufficiency when paired with a true nonfunctional allele. How do you count that? $\endgroup$
    – Bryan Krause
    Jan 27, 2023 at 20:45
  • 1
    $\begingroup$ @BryanKrause I agree with the sentiment- it's hard! This is the kind of thing that variance partitioning claims to handle, but doesn't actually. Perhaps a more viable question would be something like how frequently variable characters segregate in a true Mendelian fashion following dominance, vs. complex, vs. additive or whatever. In which case, almost none would be dominant I'd bet! But it's highly contingent on the population and whatever other variation exists. $\endgroup$ Jan 27, 2023 at 22:34

1 Answer 1

2
$\begingroup$

It is not clear whether this is an outcome of the real state of affairs, or simply that the additive model is "good enough" for summarizing the complex reality of underlying epistasis, dominance, etc (I believe "good enough", but that's somewhat philosophical).

I think it is likely that it is both. Additivity is real biologically and isn't just convenient, but it is also true that if we had complete information about every genotype-phenotype relationship, we would improve our models a bit beyond the purely additive model (for many phenotypes).

My question is this: what proportion of variants with quantifiably large phenotypic effects fit a dominance model of inheritance? Proportion of variants in the genome would be hard to define. Most nucleotides are probably useless and don't significantly affect any phenotype, drastically reducing the proportion of dominant and additive alleles. The proportion of variants that effect a given trait is easier. This is different than "importance" of dominance/additivity: you can have a lot of dominant alleles with little affect, so the phenotype is overall additive but has more dominant alleles governing it, or vice versa.

We can try to get an idea of both the raw number and the variance explained. For a raw number, and picking educational attainment as an example, Okbay et al. 2022 looked at 2,574,253 people and found that their "dominance GWAS identifies no genome-wide-significant SNPs. Moreover, with high confidence, we can rule out the existence of any common SNPs whose dominance effects explain more than a negligible fraction of the variance in EA." The "combined variance explained by dominance deviations in common SNPs is negligible." Since they found thousands of additive alleles, the proportion of dominant alleles for this phenotype is zero. For other phenotypes, I'm sure someone has found non-zero dominance alleles.

It seems like the core of what you want is how much variance is explained by additive vs. dominance/epistasis effects. I don't like this too much for the reason you stated in your comment, but it's interesting nonetheless. There have been a number of estimates on this.

A study on 70 complex traits on 254,679 individuals "found strong evidence for additive variance," but negligible dominance variance" and "epistatic variance... not significantly different from zero." "Genetic variance for complex traits is predominantly additive." "Epistatic variance is likely to be extremely small in human complex traits." This replicated a similar earlier result that looked at 79 phenotypes. This was replicated again here with 50 phenotypes. Each time, dominance is negligible.

Hill et al. "evaluate the evidence from empirical studies of genetic variance components and find that additive variance typically accounts for over half, and often close to 100%, of the total genetic variance."

Twin studies also support the additive model being predominant, with no effect of dominance for most phenotypes, but some effect for some phenotypes.

You wondered if it was the "the real state of affairs" or just "good enough." Here are some reasons why there might be real biological bias for additivity--though this isn't mutually exclusive with it being "just good enough."

  • Imagine if a gene for having a long neck was dominant in giraffes--it would be impossible to eliminate the recessive allele from the population (aside from genetic drift), as there is no phenotypic difference between heterozygous and homozygous long-neck individuals. Now imagine if there was simultaneously a different, additive pathway to having a long neck. Selection would favor increasing neck length via the additive pathway, as there would be decreased risk of one's offspring having a short neck. In this case, selection inherently favors additivity.
  • Mutations happen randomly. By definition, in order for a new dominant interaction to occur, two mutations must happen. If the mutations require each other to be useful, they'd have to occur simultaneously to be selected for. On the other hand, a new additive mutation being selected for only requires that a single new mutation is beneficial.
  • Most allele frequencies are biased towards one extreme or another. Therefore, the remaining genetic variation between individuals is likely to be of small effect and polygenic.
  • [Multilocus epistasis itself mostly contributes to additive variance.][6]
  • The existance of polygenicity alone can be sufficient for a lack of dominance effects.
$\endgroup$
5
  • 1
    $\begingroup$ I like your list of arguments about "the real state of affairs", and it seems to be a good set of arguments that dominance is very uncommon, which is great because that's the point of the question. However, it is not a good set of arguments that additivity has a real biological basis- e.g. your point 4 is explicitly a "additivity is good enough" point. Overall, thanks for the comprehensive answer! $\endgroup$ Jun 15, 2023 at 17:12
  • $\begingroup$ Right. I also agree that it is "good enough." However I do think at least points 1 and 2 are relevant to the biological basis of additivity. $\endgroup$
    – BigMistake
    Jun 15, 2023 at 17:39
  • $\begingroup$ Mm... I think that points 1+2 make arguments for the relative frequency of additivity under a toy model of selection. But they're not really biological arguments (as I usually think of them). They're saying effectively, "if additivity existed, then selection would like it." (A corollary of these is that additive variation would become fixed!) They're not saying "under the hood, there are biochemical, cytological, developmental, and ecological reasons why a certain genotype::phenotype map exists". The latter is more interesting to me, which shows my biases. $\endgroup$ Jun 15, 2023 at 18:23
  • 1
    $\begingroup$ You may be interested in this paper: academic.oup.com/genetics/article/121/4/869/5997956 $\endgroup$
    – BigMistake
    Jun 15, 2023 at 18:27
  • $\begingroup$ I don't have a citation for this, but I imagine a common biochemical reason as follows: imagine you have a typical protein-coding gene or promoter. You alter a random base pair. Odds are, the promoter or the protein is still going to function, just slightly better or worse as most amino acid sequence changes wouldn't have devastatingly large effects outside of specific critical regions (BRCA2 is a good example of this). You mostly modify the binding affinities by modifying a random base pair, which wouldn't result in dominance. If most mutations killed the gene, dominance would be more common $\endgroup$
    – BigMistake
    Jun 15, 2023 at 18:32

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .