Reading: Polygenic Inheritance and Environmental Effects

How is Height Inherited?

Many heritable human characteristics don’t seem to follow Mendelian rules in their inheritance patterns. For example, consider human height. Unlike a simple Mendelian characteristic, human height displays:

  • Continuous variation. Unlike Mendel’s pea plants, humans don’t come in two clear-cut “tall” and “short” varieties. In fact, they don’t even come in four heights, or eight, or sixteen. Instead, it’s possible to get humans of many different heights, and height can vary in increments of inches or fractions of inches. As an example, consider the bell curve-shaped graph below, which shows the heights of a group of male high school seniors.
  • A complex inheritance pattern. If you’ve paid attention to the heights of your friends and family, you may have noticed that many different patterns of inheritance are possible. Tall parents can have a short child, short parents can have a tall child, and two parents of different heights may or may not have a child of intermediate height. In addition, siblings with the same two parents may have a range of heights, ones that don’t fall into clear, distinct categories. Simple models involving one or two genes can’t accurately predict all of these inheritance patterns.
 Histogram showing height in inches of male high school seniors in a sample group. The histogram is roughly bell-shaped, with just a few individuals at the tails (60 inches and 77 inches) and many individuals in the middle, around 69 inches.

Image modified from “Continuous variation: Quantitative traits,” by J. W. Kimball (CC BY 3.0).

How, then, is height inherited? Height and other similar features are controlled not just by one gene, but rather, by multiple (often many) genes that each make a small contribution to the overall outcome. This inheritance pattern is called polygenic inheritance (poly– = many). For instance, a recent study found over 400 genes linked to variation in height[1]. When there are large numbers of genes involved, it becomes hard to distinguish the effect of each individual gene, and even harder to see that gene variants (alleles) are inherited according to Mendelian rules. In a further complication, height doesn’t just depend on genetics: it also depends a lot on environmental factors, such as a child’s overall health and the type of nutrition he or she receives while growing up.

In this article, we’ll look in more detail at how complex human traits such as height are inherited, as well as how factors like genetic background and environment can influence the phenotype (observable features) produced by a particular genotype (set of gene variants, or alleles).

Polygenic Inheritance

Some human characteristics, such as height, eye color, and hair color, don’t come in just a few distinct forms. Instead, they vary in small gradations, forming a spectrum or continuum of possible phenotypes. Features like these are called quantitative characters, and they’re typically controlled by multiple genes (often, many genes), each of which contributes to the overall phenotype. For example, although there are two major eye color genes, there are at least 14 additional genes that play roles in determining a person’s exact eye color[2].

Looking at a real example of a human polygenic trait would get complicated, largely because we’d have to keep track of tens, or even hundreds, of different allele pairs. However, we can use an example involving the color of wheat kernels to see how Mendelian inheritance of multiple genes (plus a little incomplete dominance of alleles) can produce a broad spectrum of phenotypes[3]. In this example, there are three genes that make reddish pigment in wheat kernels, which we’ll call A, B, and C. Each comes in two alleles, one of which makes a unit of pigment (the capital-letter allele) and one of which does not make any pigment (the lowercase allele). Thus, the aa genotype would contribute zero units of pigment, the Aa genotype would contribute one unit, and the AA genotype would contribute two—basically, a form of incomplete dominance.

64-square Punnett square illustrating the phenotypes of the offspring of an AaBbCc x AaBbCc cross (in which each uppercase allele contributes one unit of pigment, while each lowercase allele contributes zero units of pigment). Of the 64 squares in the chart: 1 square produces a very very dark red phenotype (six units of pigment) 6 squares produce a very dark red phenotype (five units of pigment) 15 squares produce a dark red phenotype (four units of pigment). 20 squares produce a red phenotype (three units of pigment) 15 squares produce a light red phenotype (two units of pigment) 6 squares produce a very light red phenotype (one unit of pigment) 1 square produces a white phenotype (no units of pigment)

Now, let’s imagine that two plants heterozygous for all three genes (AaBbCc) were crossed to one another (or, equivalently, allowed to self-fertilize). Each of the parent plants would have three units of pigment, or pinkish kernels. Their offspring, however, could display seven different categories of phenotypes, ranging from zero units of pigment (aabbcc) and pure white kernels to six units of pigment (AABBCC) and dark red kernels, with the intermediate phenotypes being most common.

This example illustrates how we can get a spectrum of slightly different phenotypes (something approaching continuous variation) with just three genes whose alleles display incomplete dominance. It’s not hard to imagine that, as we increased the number of genes involved, we’d be able to get even finer variations in color, or in another trait such as height. Real polygenic traits aren’t usually quite this clean and simple. (For instance, genes may make unequal contributions to the phenotype, alleles may or may not display incomplete dominance, and there may be non-additive interactions between genes.) However, the basic idea—that multiple genes obeying Mendelian rules can produce a spectrum of finely differing phenotypes—holds true for human traits such as skin and eye color.

Variable Penetrance and Incomplete Expressivity

Even for simpler characteristics that are primarily controlled by a single gene, it’s possible for individuals with the same genotype to show different phenotypes. For example, in the case of a genetic disorder, it’s possible for people with the same disease-causing genotype to have stronger or weaker forms of the disorder, or for some of them to never develop the disorder at all.

  • In variable expressivity, a phenotype varies in strength between different individuals with the same genotype. For instance, among people with the same genotype for a disease-causing gene, some might develop a severe form of the disorder, while others might have a milder form. The concept of expressivity is illustrated in the diagram below, in which the shade of green represents the strength of the phenotype.
     Narrow expressivity: all six squares are dark green. Variable expressivity: the six squares are various shades of green. The squares in each example are intended to represent individuals of the same genotype for the gene of interest.

    Illustration modeled after similar image by Steven M. Carr, Penetrance versus expressivity.

  • In incomplete penetrance, individuals with a particular genotype may or may not develop a phenotype associated with that genotype. For example, among people with the same disease-causing genotype for a particular hereditary disorder, some might never actually develop the disorder. The concept of penetrance is illustrated in the diagram below, in which green or white color represents presence or absence of a phenotype.
     Complete penetrance: all six squares are dark green. Incomplete penetrance: three of the squares are dark green, and three of the squares are white. The squares in each example are intended to represent individuals of the same genotype for the gene of interest.

    Illustration modeled after similar image by Steven M. Carr, Penetrance versus expressivity.

What causes variable expressivity and incomplete penetrance? While a variety of factors may contribute, two important ones appeared in earlier sections of this article: other genes and environmental effects. For example, disease-causing alleles of a particular gene might be suppressed by alleles of another gene elsewhere in the genome, such that individuals with the suppressor alleles have a mild form of a disorder (or even fail to develop the disorder at all). Similarly, environmental factors—such as exposure to toxins or chemicals, or overall physical health—might affect whether a person expresses a genetic disorder, as well as the severity of the disorder. A person’s sex, as well as the source of a disease-causing allele (whether it comes from the mother or father), may also affect expressivity and penetrance in some cases.


  1. Wood, A. R., Esko, T., Yang, J., Vedantam, S., Pers, T. H., Gustafsson, S., ... Frayling, T. M. (2014). Defining the role of common variation in the genomic and biological architecture of adult human height. Nature Genetics, 46, 1173–1186. http://dx.doi.org/10.1038/ng.3097.
  2. White, D. and Rabago-Smith, M. (2011). Genotype-phenotype associations and human eye color. Journal of Human Genetics, 56, 5–7. http://dx.doi.org/10.1038/jhg.2010.126.
  3. Kimball, J. W. (2011, March 8). Continuous variation: Quantitative traits. Retrieved from http://users.rcn.com/jkimball.ma.ultranet/BiologyPages/Q/QTL.html.