DNA, Genes and Chromosomes: How the Pieces Fit
Two metres of DNA in every cell, 20,000 genes, and the 98% that does not code for protein.
The words get used loosely and they describe things at very different scales. Getting the nesting straight makes everything downstream easier to read.
The nesting
- DNA — the molecule. Two strands twisted together, each a sequence of four bases: A, C, G and T. A pairs with T, C pairs with G, so each strand carries the complete information and can rebuild the other.
- Gene — a stretch of DNA with a job, usually the instructions for building one protein.
- Chromosome — one very long DNA molecule wound around proteins. Humans have 23 pairs: 22 numbered, plus the sex chromosomes.
- Genome — the whole set. About 3.1 billion base pairs.
The packing is the remarkable part. Stretched out, the DNA in a single cell is roughly two metres long, and it fits inside a nucleus about six micrometres across — wound around histone proteins into nucleosomes, coiled, and coiled again.
Two copies of everything
Chromosomes come in pairs, one from each parent, so you carry two copies of nearly every gene. The copies can differ, and those variants are called alleles. Whether a variant shows up in you depends on how the two copies interact — the subject of genotype and phenotype.
The exception is the sex chromosomes. Someone with one X has a single copy of everything on it, which is why X-linked conditions show a distinctive inheritance pattern.
Only about 2% codes for protein
Roughly 20,000 protein-coding genes occupy a small fraction of the genome. The rest was once labelled junk DNA, and that has not aged well — a great deal of it does identifiable work:
- Regulatory sequences — promoters and enhancers that determine when and where a gene is switched on. A liver cell and a neuron carry identical DNA and differ almost entirely in what is switched on.
- Non-coding RNA genes — transcribed into RNA that functions as RNA rather than being translated.
- Structural sequence — centromeres and telomeres, which keep chromosomes intact and correctly divided.
- Transposable elements — repeated sequences making up nearly half the genome, mostly inert, occasionally co-opted.
This is why "the gene for" a trait is almost always wrong. Most traits involve many genes plus the regulation determining how much of each protein is made.
From gene to protein
Transcription copies a gene into messenger RNA. Splicing removes the non-coding introns and joins the coding exons. Translation reads the result three bases at a time — each triplet specifying one amino acid — and the chain folds into a protein.
Alternative splicing means one gene can yield several different proteins by joining its exons in different combinations, which is a large part of how 20,000 genes support far more than 20,000 functions.
Where consumer tests fit
A consumer DNA test does not read your genome. It genotypes several hundred thousand specific positions known to vary between people — well under 0.03% of the sequence, chosen because those positions are informative. That is a different thing from sequencing, and it is the key to reading a raw data file without over-reading it.
Genotype and phenotype · SNPs explained · Back to all articles