Lesson 19: Genetics in the Real World — Disorders, Testing, and Gene Therapy

Genetics meets the clinic

Lessons 3–5 built the tools: inheritance patterns, DNA tests, engineered genes. This lesson uses them on real human conditions — how genetic disorders are diagnosed, how risk is calculated, and how the newest treatments rewrite the gene itself.

Patterns of genetic disorder

Two inheritance subtleties examiners love: dominant disorders can appear in every generation (Huntington's: an affected heterozygote passes it to ~50% of children), and late onset lets dominant disorders evade selection — Huntington's symptoms typically begin at 35–45, after most people have had children, so the allele persists despite being lethal in effect. Cystic fibrosis's persistence is the opposite story: carriers of many recessive alleles are invisible to selection, so the allele hides in the population (2pq can be large while q² stays small — Lesson 6's maths again).

Worked example: carrier frequency from incidence

Cystic fibrosis affects about 1 in 2,500 European-ancestry newborns. q² = 1/2,500 = 0.0004 → q = 0.02, p ≈ 0.98. Carrier frequency = 2pq = 2 × 0.98 × 0.02 ≈ 0.039 ≈ 1 in 25 people is a carrier — most with no family history of the disease, which is why couple-based carrier screening (offered before or during pregnancy) finds at-risk couples that pedigrees never would. If both partners are carriers, each child faces the 1-in-4 risk of Lesson 3's Punnett square.

Genetic testing: kinds and questions

• Diagnostic testing — confirms a suspected condition from symptoms.
• Predictive/presymptomatic testing — e.g. Huntington's: near-certain knowledge of a future untreatable disease; offered only with extensive counselling, and adults choose for themselves whether to know.
• Carrier screening — recessive risk for couples.
• Prenatal testing — screening (blood tests, ultrasound, and NIPT — cell-free fetal DNA in maternal blood, testing from a blood draw) versus diagnosis (amniocentesis/CVS, sampling fetal cells with a small miscarriage risk) — the screening/diagnosis distinction: screening estimates risk; only diagnosis confirms.
• Newborn screening — heel-prick blood spot for treatable conditions (PKU, Lesson 3) where early action prevents damage.

Testing raises issues beyond accuracy: who else does a result implicate (a positive test diagnoses your siblings' risk too), who may see it (insurers, employers — many countries restrict genetic discrimination by law), and incidental findings (searching for one gene reveals another risk).

Gene therapy: fixing the message

Gene therapy delivers a working gene to a patient's cells. *Ex vivo*: cells are removed, corrected in the lab, and returned — used for blood disorders; the 2023-approved CRISPR treatment for sickle cell disease edits patients' own stem cells to switch fetal haemoglobin back on, compensating for the faulty adult gene. *In vivo*: a vector (often a modified virus, engineered not to cause disease) delivers the gene directly — used for an inherited blindness (Luxturna) and spinal muscular atrophy. Haemophilia gene therapy can free patients from weekly factor injections for years. Limits are real: delivery to enough cells, immune reactions to vectors, durability, cost (some therapies exceed $2 million per patient — an access and equity question for Lesson 20's ethics), and the absolute rule in most jurisdictions: somatic (body-cell) therapy only, never heritable germline editing.

Pharmacogenomics: the right drug at the right dose

Genes also decide drug response: liver-enzyme variants make some people metabolise codeine dangerously fast (converting it to morphine) and others not at all; a gene variant predicts severe reactions to the HIV drug abacavir, so patients are tested before prescription. Dosing by genotype — pharmacogenomics — is genetics quietly entering everyday prescribing.

Common misconceptions

• A positive predictive test is not always a diagnosis of present disease. For Huntington's it forecasts onset with near certainty; for cancer-risk genes (BRCA1/2) it gives *probabilities* (elevated lifetime risk, not certainty) — risk literacy matters.
• Gene therapy does not (legally) change the patient's children. Somatic edits are not in gametes — Lesson 5's distinction, with regulatory teeth.
• "The gene for X" is usually shorthand. Most traits and common diseases are polygenic and environmental; single-gene disorders are the clear, teachable minority.

Exam Experience

Question 1. A genetic disorder appears in every generation of a family, and affected individuals have one affected parent. The disorder is most likely —

• A. autosomal recessive
• B. autosomal dominant
• C. caused by a recessive allele carried silently
• D. non-genetic

Question 2. If a disease affects 1 in 10,000 people and is autosomal recessive, the carrier frequency is approximately —

• A. 1 in 10,000
• B. 1 in 100
• C. 1 in 50
• D. 1 in 5

Question 3. Gene therapy for a blood disorder that removes stem cells, corrects them in the laboratory and returns them is called —

• A. germline therapy
• B. ex vivo somatic therapy
• C. cloning
• D. prenatal screening

Answers. 1: B. 2: C — q² = 0.0001, q = 0.01, 2pq ≈ 0.0198 ≈ 1 in 50. 3: B.

Constructed response. A couple, both unaffected, have a child with an autosomal recessive disorder. Neither family has any history of the condition. (a) Explain how two unaffected parents can have an affected child (2 points). (b) State the probability that their next child is a carrier (1 point). (c) Explain why "no family history" is expected for recessive conditions, using allele frequencies (2 points). Scoring guide. (a) Both parents are heterozygous carriers (1 pt); the child inherited the recessive allele from each (1 pt). (b) 1 in 2 (50%) — of the Punnett outcomes FF, Ff, Ff, ff, excluding the affected ff gives 2/3 among unaffected, but the standard school answer asked for probability at conception: accept 50% (carrier outcomes Ff = 2 of 4) or the conditional 2/3 with explanation (1 pt). (c) Recessive alleles persist hidden in carriers (1 pt); carriers (2pq) vastly outnumber affected individuals (q²) when q is small (1 pt).

Going deeper: epigenetics — inheritance beyond the sequence

Epigenetics studies gene-activity patterns that are set without changing the DNA sequence — chemical tags (methylation of DNA, modifications of the histone proteins DNA wraps around) that switch genes off or tune them, and can persist through cell divisions. It explains how identical DNA produces a neuron and a hepatocyte (Lesson 15), how identical twins drift apart molecularly with age, and some environmental echoes: children of mothers who experienced famine during pregnancy (the Dutch Hunger Winter cohort) show altered methylation at metabolic genes and different health profiles decades later. Handle the topic with care: most epigenetic marks are reset between generations, transgenerational claims are actively researched rather than settled, and textbooks that announce "Lamarck was right" are overstating. The measured version: development is sequence *plus* state, and state is partly environmental.

Going deeper: polygenic risk and the limits of prediction

Most common diseases — heart disease, type 2 diabetes, schizophrenia — involve hundreds to thousands of genetic variants of tiny effect, summed by polygenic risk scores. The top few percent of a score distribution can carry risk comparable to some single-gene disorders, which tempts clinics and direct-to-consumer companies alike. But scores predict best in the ancestry they were trained on (mostly European datasets — an equity problem in the data itself), explain only part of risk (environment and chance remain large), and describe populations better than individuals: a "high" score usually means a modest absolute risk increase. Genetic literacy for citizens now includes asking three questions of any test: How big is the effect? In whom was it validated? What would I do differently with the answer?

Worked example 2: BRCA — risk, honestly stated

A pathogenic BRCA1 variant raises lifetime breast-cancer risk from ~12% (general population) to roughly 55–70%, and ovarian risk from ~1.5% to ~40%. Note the form of the information: elevated probability, not destiny — some carriers never develop cancer, and most breast cancers occur in non-carriers. Decisions that follow (enhanced screening, risk-reducing surgery, family notification) are values-laden judgements made with counsellors, not calculations with one right answer — the human core of clinical genetics.

In practice: the counselling room

Genetic counselling is where this lesson's numbers become conversations. A counsellor's craft is non-directive: present the risk in multiple formats (1 in 4, 25%, "of 100 couples like you, about 25…"), check understanding, surface the options — conceive naturally with prenatal testing, IVF with embryo testing (PGT), donor gametes, adoption, remaining childfree — and support the couple's own values rather than steering. The same test result leads different families to opposite decisions, both legitimate. This is also why access matters: counselling quality cannot be a luxury good when the information itself is becoming cheap. The science hands people probabilities; the humanity is in what happens next.

Final check: pedigree patterns in ten seconds

Examiners show pedigrees; pattern recognition scores them. Autosomal dominant: appears in every generation; affected individuals usually have an affected parent; both sexes equally; two affected parents can have an unaffected child (both heterozygous). Autosomal recessive: skips generations; unaffected carrier parents have affected children; both sexes equally; more common when parents are related (consanguinity raises the chance both carry the same rare allele — a frequent real-world context). X-linked recessive: mostly males affected; carrier females pass it to sons; an affected male cannot pass it to his sons (they get his Y) but all his daughters are carriers. X-linked dominant: affected males pass it to all daughters and no sons. State the pattern, cite the clue (a skipping generation, an affected daughter of an unaffected father excludes X-linked recessive), then calculate. Three moves, in that order, answer almost every pedigree question set.