AP Syllabus focus:
‘Some traits show phenotypic ratios that differ from Mendel’s predictions, revealed through quantitative analysis of offspring.’
Mendelian genetics provides clear expected ratios for simple traits, but real inheritance data often deviate. AP Biology emphasises recognising these deviations and using quantitative comparisons of observed versus expected offspring outcomes.
Mendelian ratios and what they assume
Mendel’s classic ratios (for example, 3:1 phenotypes in a monohybrid cross) arise only when a trait fits a simplified genetic model.
Core assumptions behind Mendelian predictions
A single gene controls the trait.
There are two alleles in the cross.
Genotypes map cleanly to phenotypes (often taught as complete dominance).
Each offspring outcome is equally likely given parental gametes.
All genotypes have equal viability (no genotype is missing because it is lethal or severely reduces survival).
Offspring counts are large enough that chance variation is small.
When one or more assumptions fail, the phenotypic ratio you count in offspring can differ from the expected Mendelian ratio.
Quantitative analysis: comparing expected vs observed
Deviations are identified by collecting offspring data and comparing it to a predicted ratio from a genetic hypothesis.
Expected ratio: the offspring phenotypic (or genotypic) proportions predicted by a stated inheritance model, used as a baseline for comparison to real data.
Observed results rarely match expected values perfectly, even when the model is correct, due to random chance in fertilisation and survival. Quantitative analysis helps decide whether differences are small enough to attribute to sampling variation.
Chi-square as a tool for evaluating deviation
A common statistical approach is the chi-square test, which evaluates whether observed counts differ from expected counts more than would be likely by chance.
Chi-square test (): a statistical test that compares observed and expected categorical outcomes to evaluate whether deviations are consistent with random sampling under a null hypothesis.
In AP Biology, the interpretation matters more than computation: a small suggests the model may fit; a large suggests the model may not fit.
= Observed count in a category (offspring)
= Expected count in a category (offspring)
= Degrees of freedom (number of categories )
To interpret results, students typically use a p-value threshold (often 0.05) to decide whether to reject the model.
This figure shows a right-skewed chi-square distribution with a test statistic marked on the x-axis and the right-tail region shaded to represent the p-value. It helps connect the computed \chi^2$ value to the probability of seeing deviations at least this large under the null hypothesis (i.e., if the genetic model is correct). Source
This decision always depends on assumptions such as independent observations and appropriate sample size.
Why traits deviate from Mendelian ratios
Deviations often reflect biology that is more complex than “one gene, two alleles, complete dominance.”
Allele-to-phenotype relationships differ from simple dominance
Some traits do not produce just “dominant” and “recessive” phenotypes. When heterozygotes have their own phenotype category, the expected phenotypic ratio changes because genotype categories no longer collapse into two phenotype groups.
This figure illustrates incomplete dominance in snapdragon flower color across P, F1, and F2 generations. Because the heterozygote has an intermediate phenotype, the phenotypic ratio in the F2 matches the genotypic ratio (1:2:1) rather than collapsing into a 3:1 pattern. Source
Key idea: more phenotype categories can appear, shifting predicted ratios away from 3:1.
More than two alleles may exist in a population
Even if an individual carries only two alleles, a gene can have multiple alleles in the population. This expands possible genotype combinations and may yield more complex phenotype groupings than Mendel’s simplest crosses.
Key idea: allelic diversity can complicate predictions for crosses beyond the basic two-allele setup.
One gene can affect whether another gene is expressed
A phenotype may depend on interactions among genes in the same organism, so that one gene’s alleles modify or mask the phenotypic effects of another gene.
Key idea: gene interaction can collapse or redistribute phenotype categories compared with the 9:3:3:1 expectation from two independently acting genes.
Some genotypes are missing from the offspring
If a genotype reduces survival before counting (for example, during development), observed offspring ratios can shift because not all fertilisation outcomes are represented.
Key idea: viability differences produce fewer individuals in certain categories than expected.
Traits may be controlled by many genes with small effects
When many genes contribute to a trait, phenotypes often show continuous variation rather than discrete categories, and simple Mendelian ratios are not expected.
This diagram compares frequency distributions for a trait controlled by 1 locus, 2 loci, 3 loci, and many loci. As more loci contribute additively, the number of phenotypic classes increases and the distribution approaches a smooth, continuous curve—matching how many quantitative traits are analyzed as distributions rather than as Mendelian ratios. Source
Key idea: polygenic inheritance changes the kind of data you analyse (often distributions rather than simple category ratios).
Random sampling effects (especially with small sample sizes)
Even with a correct model, small broods can deviate substantially from expected ratios just by chance.
Key idea: larger sample sizes generally produce ratios closer to expectation, making quantitative analysis more reliable.
FAQ
Use the phenotypes you can score reliably and consistently.
If phenotypes are ambiguous, combine categories only if justified by the genetic hypothesis.
Very small expected counts per category can break assumptions.
Non-independent observations (e.g., repeated measures of the same individual) also undermine the test.
Yes. The same observed ratio can be produced by different mechanisms.
Additional crosses or molecular evidence are often needed to distinguish causes.
Scoring error and misclassification of phenotypes can distort observed counts.
Unequal survival after birth but before counting can also bias the dataset.
They often measure trait values across many offspring and analyse patterns of variation.
Approaches include comparing distributions, estimating heritability, or mapping contributing loci with appropriate designs.
Practice Questions
Explain why observed offspring phenotypic ratios may differ from Mendel’s predicted ratios even when the inheritance model is correct. (2 marks)
Mentions random chance/sampling variation in fertilisation and offspring counts (1)
Mentions effect of small sample size increasing deviation from expectation (1)
A student predicts a 3:1 phenotypic ratio from a monohybrid cross but observes a different ratio in the offspring. Describe how quantitative analysis can be used to evaluate whether the data support the 3:1 model, and give two biological reasons the model might be inappropriate. (5 marks)
States comparison of observed counts to expected counts under a 3:1 hypothesis (1)
Identifies use of chi-square test / calculation of from (1)
Refers to interpreting with degrees of freedom and a p-value threshold (e.g., 0.05) to decide reject/fail to reject (1)
Gives one biological reason (e.g., heterozygote has distinct phenotype; genotype-dependent viability; gene interaction; polygenic control) (1)
Gives a second, different biological reason (1)
