1 Genetic Effects

A complete understanding of genetic effects would need to answer a series of questions:

1. How important are genetic effects on human differences?
2. What kinds of action and interaction occur between gene products in the pathways between genotype and phenotype?
3. Are the genetic effects on a trait consistent across sexes?
4. Are there some genes that have particularly outstanding effects when compared to others?
5. Whereabouts on the human gene map are these genes located?

Questions 4 and 5 are clearly very important, but are not the immediate concern of this book. On the other hand, we shall have a lot to say about 1, 2, and 3. It is arguable that we shall not be able to understand 4 and 5 adequately, if we do not have a proper appreciation of these other issues. The importance of genes is often expressed relative to all the causes of variation, both genetic and environmental. The proportion of variation associated with genetic effects is termed the broad heritability. However, the complete analysis of genetic factors does not end here because, as countless experiments in plant and animal genetics have shown (well in advance of modern molecular genetics; see e.g., Mather and Jinks, 1982), genes can act and interact in a variety of ways before their effects on the phenotype appear. Geneticists typically distinguish between additive and non-additive genetic effects (these terms will be defined more explicitly in Chapter 3). These influences have been studied in detail in many non-human species using selective breeding experiments, which directly alter the frequencies of particular genotypes. In such experiments, the bulk of genetic variation is usually explained by additive genetic effects. However, careful studies have shown two general types of non-additivity that may be important, especially in traits that have been subject to strong directional selection in the wild. The two main types of genetic non-additivity are dominance and epistasis. The term dominance derives initially from Mendel's classical experiments in which it was shown that the progeny of a cross between two pure breeding lines often resembled one parent more than the other. That is, an individual who carries different alleles at a locus (the heterozygote) is not exactly intermediate in expression between individuals who are pure breeding (homozygous) for the two alleles. While dominance describes the interaction between alleles at the same locus, epistasis describes the interaction between alleles at different loci. Epistasis is said to occur whenever the effects of one gene on individual differences depend on which genotype is expressed at another locus. For example, suppose that at locus $A/a$ individuals may have genotype $AA$, $Aa$ or $aa$, and at locus $B/b$ genotype $BB$, $Bb$ or $bb$. If the difference between individuals with genotype $AA$ and those with genotype $aa$ depends on whether they are $BB$ or $bb$, then there would be additive $\times$ additive epistatic interactions. Experimental studies have shown a rich variety of possible epistatic interactions depending on the number and effects of the interacting loci. However, their detailed resolution in humans is virtually impossible unless we are fortunate enough to be examining a trait which is influenced by a small number of known genetic loci. Therefore we acknowledge their conceptual importance and model them if they are identified. Failure to take non-additive genetic effects into account may be one of the main reasons studies of twins give different heritability estimates from studies of adoptees and nuclear families (Eaves et al., 1992; Plomin et al., 1991). As studies in genetic epidemiology become larger and better designed, it is becoming increasingly clear that there are marked sex differences in gene expression. An important factor in establishing this view has been the incorporation of unlike-sex twin pairs in twin studies (Eaves et al., 1990). However, comparison of statistics derived from any relationship of individuals of unlike sex with those of like sex would yield a similar conclusion (see Chapter 9). We shall make an important distinction between two types of sex-limited gene expression. In the simpler case, the same genes affect both males and females, but their effects are consistent across sexes and differ only by some constant multiple over all the loci involved. We shall refer to this type of effect as scalar sex-limitation. In other cases, however, we shall discover that genetic effects in one sex are not just a constant multiple of their effects in the other. Indeed, even though a trait may be measured in exactly the same way in males and females, it may turn out that quite different genes control its expression in the two sexes. A classic example would be the case of chest-girth since at least some of the variation expressed in females may be due to loci that, while still present in males, are only expressed in females. In this case we shall speak of non-scalar sex-limitation. None of us likes the term very much, but until someone suggests something better we shall continue to use it!