The same genes, different outcomes
A seed's genotype sets its potential, but the soil, water, and light decide what actually grows. Gene–environment interaction means the effect of a genotype depends on the environment — and the effect of an environment depends on the genotype. Two plants with identical genes can end up very different in rich versus poor soil, and two different genotypes can even swap ranks when the environment changes.
Geneticists capture this with a norm of reaction: a curve showing how one genotype's phenotype changes across a range of environments. If the curves for different genotypes are parallel, environment just shifts everyone equally. If they cross, the “best” genotype depends entirely on the setting — a vivid reminder that genes and environment are partners, not rivals.
Norm of reaction: yield of two genotypes across environments yield ^ .* G1 | .* . | .* .' | .* .' o G2 | .*.' o o | *.' o o +------------------------------> environment (poor -> rich) In a poor environment G2 wins; in a rich one G1 wins. The lines CROSS -> there is no single 'better' genotype.
Twins: nature's controlled experiment
How do we estimate heritability in humans, where we can't run controlled crosses? One classic tool is the twin study. Identical (monozygotic) twins share ~100% of their genes; fraternal (dizygotic) twins share ~50%, like ordinary siblings. If a trait is strongly genetic, identical twins should match more often than fraternal twins do.
We measure “matching” with concordance: the chance that if one twin has a trait, the other does too. Higher concordance in identical than fraternal twins points to genetic influence. A rough rule estimates heritability as twice the gap in correlations between the two twin types. The catch: it assumes both kinds of twins share their environments equally, which is never perfectly true — so twin estimates are useful, but they are estimates, not certainties.
When continuous becomes all-or-nothing
Some multifactorial conditions look binary — you either have them or you don't — yet they still arise from many genes and environmental factors. The bridge is the threshold trait model. Imagine an invisible, continuous liability (your total genetic plus environmental risk) that follows a bell curve. Below a threshold, nothing shows; cross it, and the condition appears.
Threshold (liability) model:
number of
people
^ ___
| / \ threshold
| / \ |
| / \___ v
| / unaffected | affected
+----------------------|----------> liability
|
(genes + environment push you left or right)
Relatives of an affected person sit, on average, a bit
further right -> their risk of crossing the line is higher.
This explains why such conditions 'run in families' without
following a clean Mendelian ratio.