The three question types
| Type | What it asks | How to prepare |
|---|---|---|
| Pattern recognition | Given a pedigree or ratio, which mode of inheritance? | Volume. Fifty pedigrees, mixed, until the patterns are instant. |
| Calculation | Probability, recombination frequency, allele frequency | Drill the four or five formulas until they're automatic under time. |
| Mechanism explanation | Why does this phenomenon produce this result? | Explain out loud. This is where definitions finally matter. |
Look at your past papers and count the marks in each. In most courses the first two dominate, and most students' revision is entirely aimed at the third.
Pedigrees: build the elimination routine
A pedigree question is a decision procedure, and running it the same way every time is what makes it fast and reliable under pressure. Improvising each time is how people talk themselves out of a correct answer.
- 1
Does it skip generations?
Skipping strongly suggests recessive. Appearing in every generation suggests dominant. This one question resolves half the cases before you've looked at anything else.
- 2
Are affected males and females roughly equal?
A strong male excess points at X-linked. Roughly equal points at autosomal. Count, don't eyeball — small pedigrees mislead.
- 3
Is there father-to-son transmission?
If yes, it is not X-linked, full stop. This is the single most useful eliminating fact in pedigree analysis and it's decisive on its own.
- 4
Do all children of affected mothers show it?
That pattern points at mitochondrial inheritance, which is rare in life and common in exams precisely because it's a clean discriminator.
- 5
Check consanguinity and unaffected carriers
A consanguineous marriage in the pedigree is almost always the examiner signalling recessive.
- 6
State the answer and the eliminating evidence
"Autosomal recessive — skips generations, sexes equally affected, consanguinity present." Mark schemes usually award for the reasoning, not just the conclusion.
Probability is where the marks leak
Most genetics calculation errors are probability errors rather than genetics errors. Three rules cover nearly all of them, and they're worth being able to state precisely.
- Multiply for 'and', add for 'or'. "Affected and male" multiplies; "affected or carrier" adds. Half of all mistakes are the wrong operation.
- Conditional probability changes the denominator. "Given that this child is unaffected, what's the chance they're a carrier?" is 2/3, not 1/2, and this specific question appears on almost every paper.
- Independent events stay independent. Three unaffected children don't change the fourth child's probability. Examiners test this deliberately because the intuition is so strong.
Write these three on your cheat sheet and practise them separately from genetics, on pure probability questions, until the operations are automatic. Mixing an unstable probability skill with unfamiliar genetics content is why easy marks get lost.
Molecular genetics needs a different method
The second half of most genetics courses is molecular — replication, transcription, regulation, repair, techniques — and it isn't problem-shaped in the same way. It's mechanism-shaped, and the method that works is drawing the process and then breaking it.
Draw the pathway from memory. Then, for each step, ask what happens if this enzyme is missing, this sequence is mutated, this signal is absent. Those failure questions are simultaneously the mechanism check and the disease content, and they're what the exam asks. Same approach as physiology: the loop, then the break.
The terminology, handled efficiently
Genetics vocabulary is genuinely large and genuinely arbitrary, which is exactly the profile for spaced repetition rather than for re-reading. But make the cards discriminative, not definitional — the exam rarely asks "what is penetrance", it asks you to spot incomplete penetrance in a pedigree that otherwise looks like clean dominance.
- Card the confusable pairs together: penetrance vs expressivity, epistasis vs pleiotropy, linkage vs association. Learning them separately guarantees you'll confuse them.
- Add a 'how would you spot this' side to every phenomenon card. That's the examinable form.
- Include the exceptions to Mendelian ratios by their ratio: 9:7, 12:3:1, 15:1 each signal a specific interaction, and recognising the ratio is faster than reasoning it out under time.
A revision week that fits the exam
Allocate by mark distribution, not by how much lecture time each topic got. In a typical paper that means roughly half your revision on problems, a quarter on molecular mechanisms, and a quarter on the vocabulary that supports both — which is close to the inverse of how most people revise this subject.
And do the problems mixed. A pedigree presented in the pedigree chapter has already told you it's a pedigree question; the exam won't. Interleaving matters more here than in almost any other biological subject, because pattern recognition is the whole first question type.