Appearance
2.7 — Chromosomes, Sex Determination and Linkage
Human cells were thought to contain 48 chromosomes for over thirty years. The figure came from a 1921 count and was repeated in every textbook. In 1955 Joe Hin Tjio and Albert Levan, using a better technique for spreading the chromosomes flat, counted 46 — repeatedly and unambiguously. When they went back through published photographs from the intervening decades, several clearly showed 46, and had been captioned 48.
Everyone had counted what they expected to count. It is a useful cautionary tale about looking at evidence, and it happened in a field that had no obvious reason for bias.
Packing two metres into six micrometres
Chapter 2.4 introduced the nucleosome. Here is the full ladder, with the numbers.
The bare double helix is 2 nanometres wide. DNA wraps 1.65 times around a histone spool, and the spools sit every 200 base pairs or so, giving the 11 nm beads-on-a-string. That fibre coils into a 30 nm fibre. The 30 nm fibre is thrown into loops anchored on a protein scaffold, giving a 300 nm fibre. During division those loops are compressed further into the 700 nm chromatid, and a full chromosome at metaphase is about 1,400 nm across.
Overall compaction is roughly 10,000-fold in a working nucleus and about 50,000-fold at metaphase.
Chromosomes are only visible during division. For the rest of the cell's life the DNA is spread out as chromatin, because a gene packed into a metaphase chromosome cannot be read. This is why every photograph of chromosomes is taken from a dividing cell, and why a karyotype requires growing cells in culture first.
Each chromosome occupies its own territory in the nucleus rather than floating loose and tangling with its neighbours — a fact only established with modern imaging, and one that explains why the two metres do not knot.
The karyotype
A karyotype is the full set of chromosomes photographed at metaphase and arranged in order. Yours has:
- 22 pairs of autosomes, numbered 1 to 22 roughly by decreasing size.
- 1 pair of sex chromosomes: XX in a typical female, XY in a typical male.
Total 46, written as 46,XX or 46,XY.
Each chromosome has a waist — the centromere — dividing it into a short arm (p, for petit) and a long arm (q, simply the next letter). Staining produces reproducible dark and light bands, and the numbering of those bands gives every region an address. 7q31.2 means chromosome 7, long arm, region 3, band 1, sub-band 2 — which happens to be where the cystic fibrosis gene sits.
Chromosome 1 is the largest, about 249 million base pairs and roughly 2,000 genes. Chromosome 21 is the smallest, about 47 million base pairs and around 200 to 300 genes. That size difference is why trisomy 21 is survivable and trisomy 1 is not — an extra copy of 200 genes disrupts development substantially, and an extra copy of 2,000 genes disrupts it fatally. Almost all trisomies of large chromosomes end in miscarriage, usually before the pregnancy is recognised.
Sex determination
The default developmental path is female. That is not a value judgement; it is an observation about what happens when nothing intervenes.
The Y chromosome carries a gene called SRY — sex-determining region Y — and it is a transcription factor. At about six weeks of development, the embryo has a pair of undifferentiated gonads that could become either testes or ovaries. If SRY is expressed, they become testes. If it is not, they become ovaries. Everything downstream follows from that single switch, because the testes then produce testosterone and anti-Müllerian hormone, and those two hormones drive the rest of male development. Chapter 4.5 follows the anatomy.
The Y is small and shrinking. It carries around 50 to 60 genes against the X's roughly 800 to 900, most of them concerned with male fertility. Because the X and Y are so different, they cannot recombine along their length the way a matched pair does — only two small regions at the tips pair up. Without recombination there is no way to repair damage by copying from a partner, and the Y has been losing genes over evolutionary time. Whether it will disappear is argued about; some lineages of rodents have lost it and use a different switch.
Two consequences follow that a doctor sees regularly.
Rare cases break the XX/XY rule. During meiosis in a father, the X and Y pair briefly at their tips, and occasionally the SRY gene crosses over onto the X. A child inheriting that X plus another X is 46,XX with SRY and develops as male, though infertile. The reverse — a Y that has lost SRY — gives a 46,XY female. These are rare, and they demonstrate that it is the gene rather than the chromosome that does the work.
Androgen insensitivity syndrome is the other instructive case. A person is 46,XY, has testes, and produces normal or high testosterone — but their androgen receptor is faulty, so the body cannot respond to it. Since the female path needs no signal, development follows it: the person is born with female external anatomy, is usually raised as a girl, and typically comes to attention in adolescence when periods do not start. It demonstrates the whole logic of the system: hormones are messages, and a message nobody can read has no effect.
And X-inactivation, from Chapter 2.4, is the dosage fix. A female silences one X in each cell so that the working dose matches a male's single X. Because the choice is random and made early, every female is a mosaic of cells using one X or the other. This is why a female carrier of an X-linked recessive condition is usually healthy but can show mild features if the inactivation happened to be skewed — around 10 percent of carriers of haemophilia have clotting factor levels low enough to bleed abnormally.
Linkage: the tool that mapped genes before DNA was understood
Mendel's law of independent assortment says the alleles of different genes separate independently. That is only true if they are on different chromosomes.
Two genes on the same chromosome travel together — they are linked — unless a crossover happens between them during meiosis I (Chapter 1.7). And here is the insight that built genetics as a mapping science: the further apart two genes are, the more likely a crossover falls between them.
So the frequency with which two linked traits get separated is a direct measure of the physical distance between them. Alfred Sturtevant, an undergraduate in Thomas Hunt Morgan's fruit fly laboratory, realised this in 1911 and reportedly spent a night constructing the first genetic map from existing crossing data.
Distance is measured in centimorgans: 1 cM is the distance at which a crossover occurs in 1 percent of meioses. In humans, 1 cM averages roughly 1 million base pairs, though the ratio varies a great deal along the genome because crossovers cluster in hotspots.
This is how disease genes were found before sequencing existed. You could not look at the gene, but you could look for a marker — some detectable variant — that was almost always inherited alongside the disease in affected families. Close linkage meant the gene was nearby, and then you walked along the chromosome from there. The Huntington's disease gene was located this way in 1983 and finally identified in 1993; cystic fibrosis in 1989. Both were a decade of work each, for a job that now takes a sequencing machine an afternoon.
Linkage still has direct clinical use. Genome-wide association studies work on the same principle: variants that sit close together tend to be inherited as a block, so finding a marker associated with a disease points at a nearby region even when the causal variant is not itself measured.
When chromosomes go wrong
Two categories: wrong number, and wrong structure.
Wrong number — aneuploidy. Chapter 1.7 covered nondisjunction and the resulting syndromes: trisomy 21, 18 and 13, Turner (45,X) and Klinefelter (47,XXY), with the maternal age effect explained by eggs arrested in mid-meiosis since before birth.
Add the clinical scale: chromosomal abnormalities are found in about 50 percent of first-trimester miscarriages. Most conceptions with the wrong chromosome number never implant or are lost very early. What reaches birth is the small survivable subset.
Wrong structure.
Deletion — a piece lost. Cri-du-chat syndrome comes from a deletion on the short arm of chromosome 5; affected infants have a characteristic high-pitched cry, from which the name comes, along with intellectual disability. DiGeorge syndrome (22q11.2 deletion) affects roughly 1 in 4,000 births and causes heart defects, absent or small thymus with immune deficiency, low calcium from missing parathyroid glands, cleft palate and learning difficulties — a strikingly varied list that makes sense once you know the deleted region controls development of a specific set of embryonic structures (Chapter 4.5).
Duplication — a piece present twice. Charcot–Marie–Tooth disease type 1A, the commonest inherited neuropathy, is caused by a duplication rather than a mutation: the gene is normal, there is simply too much of its product.
Inversion — a piece flipped end for end. Often harmless in the carrier, but it causes problems at meiosis because the flipped region cannot line up with its partner, so gametes come out unbalanced. A carrier is healthy and has a raised miscarriage risk.
Translocation — a piece moved to a different chromosome. This one deserves detail, because it has two very different clinical faces.
Translocations: the balanced kind, and the cancer kind
A balanced translocation means two chromosomes have swapped segments with nothing gained or lost. The carrier is completely healthy — all their genes are present, just filed in the wrong place. About 1 in 500 people carries one and most never find out.
The problem appears at meiosis. Chromosomes must pair with their partners, and a translocated chromosome can only pair awkwardly, so many gametes end up with too much of one region and too little of another. The typical consequence is recurrent miscarriage. This is why a couple with several unexplained pregnancy losses are offered karyotyping of both partners — a balanced translocation in one of them is one of the few findings that genuinely changes management.
The Philadelphia chromosome is the other face and one of the most important discoveries in cancer medicine. A piece of chromosome 9 and a piece of chromosome 22 swap, and the break points fall inside two genes — ABL on 9 and BCR on 22. The two halves are joined into a single fused gene, BCR-ABL, whose product is a permanently switched-on version of a growth-signalling kinase. The cell divides continuously, and the result is chronic myeloid leukaemia.
Imatinib was designed to fit that specific fused protein's active site. Before it, chronic myeloid leukaemia had a median survival of about five years. With it, ten-year survival is over 80 percent and life expectancy for a patient responding well is close to normal. It is the founding example of targeted cancer therapy, and the whole chain — a visible chromosome swap, a named fusion gene, a drug shaped to fit its product — is why this Part matters clinically rather than only theoretically. Chapter 19.7.
How chromosomes are examined today
Karyotyping is still used and still the only method that sees a balanced translocation, because nothing is gained or lost for a sequence-based method to detect. It needs dividing cells and takes days to weeks, and it resolves changes down to about 5 to 10 million base pairs.
FISH attaches a fluorescent probe to a specific sequence, so a named deletion or duplication can be checked quickly, including on non-dividing cells.
Chromosomal microarray compares the amount of DNA at thousands of positions against a reference, detecting gains and losses down to tens of thousands of base pairs — far finer than a karyotype. It is now first-line for a child with unexplained developmental delay. It cannot see balanced rearrangements.
Non-invasive prenatal testing analyses the fragments of fetal DNA circulating in the mother's blood from about ten weeks. For Down syndrome it detects over 99 percent of cases. But it is a screening test and must be described as one. In a low-risk 25-year-old, where trisomy 21 is genuinely rare, a substantial minority of positive results are false — the test's accuracy has not changed, but the proportion of positives that are real depends on how common the condition is in the group being tested. That is Bayes' theorem (Volume II, Chapter 7) applied to a real decision. A positive result must be confirmed by amniocentesis or chorionic villus sampling before any irreversible decision is made, and failing to explain that has caused real harm.
What the next page fixes
You now have the machinery, the code, the copying, the regulation, the mutations and the inheritance patterns. Chapter 2.8 puts them together on real diseases: exactly what happens in a body carrying one wrong letter, for sickle cell, thalassaemia, cystic fibrosis, haemophilia and Huntington's — from the base change through to the symptom, the treatment, and what is now curable that was not five years ago.