S6-1

Gene, genome, chromosome, and how DNA is packaged

14 min

By now you can read the machinery. You know a gene gets transcribed into RNA (S5.1) and translated into protein (S5.5), and you have watched a reading frame break under your hands. This lesson zooms out from the single gene to the whole storage system. The question is no longer "how is one message read" but "how is the entire codebase organized, named, and physically stored," and the answer runs into a problem of pure logistics: the full instruction set is about two meters of molecule, and it has to fit inside a compartment you could not see without a microscope. How life solves that packing problem shapes almost everything about how genes get switched on and off, which is where module S8 will pick up.

A gene is a stretch, a genome is the whole thing

Start with the two words people mix up most. A gene is a specific stretch of DNA that specifies one product. Usually that product is a protein, the gene is transcribed to mRNA and translated to a chain of amino acids, exactly the path you already traced. Sometimes the product is not a protein at all but a functional RNA molecule that does its job as RNA and is never translated, for example the tRNA adapters from S5.5 or the RNA parts of the ribosome. So the honest definition is broader than "a gene makes a protein." A gene is a stretch of DNA that specifies a product, and that product is usually but not always a protein.

The genome is the complete set of DNA an organism carries, every gene plus everything in between. If a gene is one function, the genome is the entire repository, source files, config, dead code, comments, and all. You do not have thousands of separate genome copies floating around, you have one genome, and copies of individual genes are read from it on demand.

The analogy: genome is the repo, gene is a module

Hold this mapping through the whole lesson. The genome is the whole repository. A gene is one function or module inside it. A chromosome, coming up next, is a large source file that groups many genes together. Alleles, also coming up, are two versions of the same module merged in from two parents. And histones, the packing proteins, act like compression plus access control: a tightly packed region is effectively unreadable until something unpacks it. Where the analogy misleads: a repo has a filesystem with names and paths, so you can jump straight to a function. The genome has no directory and no filenames. A gene is located by its physical position along the DNA and by the signals in the sequence itself, not by a lookup path. There is no index file. The cell finds a gene the way you would find a substring in one enormous unbroken text, not the way you open a named file.

Two meters of DNA in a nucleus you cannot see

Now the logistics problem, stated with numbers. The DNA in a single human cell, stretched end to end, is roughly two meters long. The nucleus that holds it is a few millionths of a meter across. That is like stuffing a strand as long as a person into a container smaller than a grain of dust, and doing it in a way that still lets you find and read any one gene on demand without tangling the whole thing into a useless knot.

You cannot just cram it in. A random wad of two-meter thread would snarl instantly and nothing could be read. The packing has to be ordered, reversible, and selective, so that the regions being used right now are accessible while the rest stays compact. That triple requirement, compact, ordered, and locally openable, is exactly what the histone system delivers.

DNA wound around histones

The first level of packing is beautifully simple. DNA is wound around small proteins called histones, which clump together into a barrel-shaped spool. Picture thread wrapped around that spool. A short stretch of the double helix loops a little under two turns around a cluster of histone proteins, and that bead of DNA plus histone core is called a nucleosome. String many nucleosomes along the DNA and you get a "beads on a string" arrangement, the DNA is the string, the nucleosomes are the beads.

Winding onto spools already shortens and tidies the molecule dramatically. But nucleosomes are only the first fold. The beaded string coils on itself into a thicker fiber, that fiber loops and folds again, and at the highest level of compaction the whole thing condenses into the dense, X-shaped bodies you may have seen in a textbook, the chromosomes. A chromosome is DNA packaged with its histones and folded up, not a different kind of molecule. It is the same double helix from S3, wound and wound until a file's worth of genes becomes a compact, movable unit.

Here is the diagram of the underlying helix again, the molecule that all this packing acts on. Rotate and inspect it, and keep in mind that everything in this lesson is about how this two-meter strand gets folded small enough to store while staying readable.

That readability point is not a detail, it is the whole reason packing matters for the rest of the course. A gene wound tightly into dense chromatin (the DNA-plus-histone packing material) cannot be transcribed, the machinery from S5.1 physically cannot get at it. A gene in a loosely packed region is available to be read. So how tightly a given stretch is packed becomes a way to control which genes are on and which are off, without changing a single base of the sequence. Hold that thought. Module S8 is built entirely on it.

The human numbers: 46 chromosomes, 23 pairs, two copies of most genes

Now the specifics for our species. A human cell holds its genome across 46 chromosomes. They are not 46 unrelated files. They come as 23 pairs, and the pairing has a clear origin: one chromosome of each pair came from one parent and the matching one came from the other parent. You inherit a full set of 23 from each side, and together they make 23 pairs, 46 total.

Because the two chromosomes in a pair carry the same genes in the same order, most of your genes come in two copies, one on each member of the pair. The two copies of a given gene are called alleles. They can be identical, or they can differ slightly in sequence, and that difference is where a lot of biology lives. One allele might carry a typo that breaks the protein while the other works fine, in which case the working copy can often cover for the broken one. Genetics, which S6 and beyond will develop, is largely the study of how these two copies interact.

In the repo analogy, alleles are two versions of the same module, one merged in from each parent, sitting side by side in the codebase. Where that picture strains: a real merge would force you to reconcile the two versions into one. The cell does not reconcile. It keeps both copies intact and, for most genes, can read either or both. Two live versions of the same module coexist rather than being flattened into one.

About 3 billion base pairs, roughly 20,000 protein-coding genes

Two headline sizes anchor the human genome. The total length is about 3 billion base pairs, three billion rungs of the A-T and G-C ladder from S3, spread across those 46 chromosomes. And the number of protein-coding genes is roughly 20,000.

Sit with how those two numbers relate, because the ratio is where the surprise is. Twenty thousand genes across three billion base pairs is not very dense. Do the rough division and the coding genes, even generously sized, do not come close to filling the space. Which raises the obvious programmer's question: if only twenty thousand genes are in three billion bases, what is all the rest of it?

The shock: only 1 to 2 percent codes for protein

Here is the fact that reliably stuns people the first time. Of the roughly three billion base pairs in your genome, only about 1 to 2 percent actually codes for protein. Say it plainly: about ninety-eight percent of your DNA is not protein-coding sequence. The recipes that get translated into the amino acid chains you have spent module S5 on make up a thin slice of the whole.

So what is the other ninety-eight percent doing? Several honest categories, in rough terms. A large fraction is regulatory, sequences that do not code a protein but control when and how much nearby genes are read, the switches and dials of module S8. Some is structural, needed for the chromosome to fold, copy, and hold together. A great deal is repetitive, the same short patterns duplicated over and over, some of it descended from ancient viruses that inserted themselves and never left. And a real, honestly acknowledged portion we simply do not yet fully understand.

Junk DNA was an overconfident label

For decades the non-coding majority was dismissed as "junk DNA," on the reasoning that if it does not code a protein it must be useless filler. That label was overconfident, and it aged badly. A lot of what was called junk turned out to be regulatory sequence doing essential control work, or structural sequence, or elements with functions we are still uncovering. That does not flip to the opposite overclaim either. It is not the case that every base is now known to be vital, plenty of the genome does look like genuine leftover repetition and inserted parasites with no current job. The honest position, and the one this course takes, is in between: non-coding does not mean non-functional, function has to be demonstrated rather than assumed either way, and "junk" was a guess dressed up as a conclusion.

Forward link to bioinformatics: the coding part is the small part

When you reach the bioinformatics track and open a real human genome, this ratio is the first thing that reshapes your intuition. You are handed three billion letters, and the genes you are hunting for occupy only one or two percent of them. Finding the coding regions inside that vast non-coding sea is exactly the ORF-scanning and gene-prediction problem from S5.5, now at genome scale, and it is hard precisely because the signal is sparse. Just as important, much of the interesting biology, the disease-linked variation, the control logic, lives out in the ninety-eight percent, in the regulatory and repetitive regions. A tool that only looked at protein-coding sequence would be blind to most of the genome. The track will make this precise. For now, absorb the shape of it: the code you know how to read is a minority of what is stored.

Key terms

gene: A stretch of DNA that specifies one product, usually a protein but sometimes a functional RNA.
genome: The complete set of DNA an organism carries, every gene plus all the sequence in between.
histone: A spool-like protein that DNA winds around, the first level of packing that helps meters of DNA fit in a nucleus.
nucleosome: One bead of the beads-on-a-string arrangement, a segment of DNA wound around a histone core.
chromosome: A single continuous DNA molecule packaged with its histones and folded, not a different substance from the DNA itself.
allele: One of the two copies of a gene, one inherited from each parent, which may be identical or differ slightly in sequence.
base pair: One rung of the DNA ladder, an A-T or G-C pairing. The human genome holds about 3 billion of them.
protein-coding sequence: The roughly 1 to 2 percent of the human genome that is translated into protein, the rest being regulatory, structural, repetitive, or not yet understood.

Why 'about 20,000 genes' is a moving, contested number

The figure of roughly 20,000 protein-coding genes sounds precise, but it hides a real and ongoing argument. Before the human genome was sequenced, many biologists bet the number would be 100,000 or more, on the intuition that a human surely needs far more genes than a worm. The actual count came in shockingly lower, near 20,000, close to what much simpler organisms carry, and that humbling result forced the field to accept that complexity comes more from how genes are regulated and combined than from sheer gene count. Even now the exact number is not settled, because "gene" is fuzzier at the edges than the tidy definition suggests. Some stretches produce several different proteins from one region, some produce only RNA and whether to count them varies, and new candidate genes are still being confirmed or discarded. So treat 20,000 as a well-supported estimate of order, not a counted constant. The uncertainty is not sloppiness, it is a sign that the definition of a gene is doing more work than it looks.

Check yourself

1. What is the difference between a gene and a genome?

2. Roughly two meters of DNA fits inside a microscopic nucleus mainly because

3. Why do most human genes come in two copies called alleles?

4. A colleague says every base of your 3 billion base pair genome must code for a protein, otherwise it would have been deleted by evolution. What is the accurate correction?

4 unanswered