Turning a gene into its corresponding protein is a two step-process. First, the gene on a piece of DNA in a chromosome is used as a template for building a molecule of another kind of nucleic acid, called RNA. This is called transcription. The piece of RNA made from a gene is then used as a template for putting the protein together, one amino acid at a time. This is called translation, and it is performed by a complex piece of molecular machinery called the ribosome, made of proteins and other pieces of RNA.
Chromosomes consist of lengths of DNA double-helix wound around disk-like protein molecules called histones, like the string on a yoyo. This combination of DNA and its protein packaging is what we call chromatin. The genomes of eukaryotes are divided up into a number of chromosomes that is always the same for every cell of a particular species (if they are not abnormal) but can differ between species. Human cells have 46 chromosomes, in sets of 23 pairs.
* * *
It’s common to see genes called the instructions to make an organism. In this view, the entire genome is then the “instruction booklet”, or even the “blueprint”. This is an understandable metaphor, but misleading. Genes are fundamental to the way an organism turns out: the genome of a frog egg guides it to become a frog, not an elephant, and vice versa. But the way genes influence and to some degree dictate that proliferation of cells is subtle, complex, and resistant to any convenient metaphors from the technological world of design and construction. By leaping from genome to finished organism without taking into account the process of development from cells, we risk simplifying biology in ways that can create some deep misconceptions about how life proceeds and evolves.
To the extent that a gene is an “instruction”, it is an instruction to build a protein molecule. It is far from obvious what, in general, this has to do with the growth and form of an organism: with the generation of our flesh. We know of no way to map an organism’s complement of proteins onto its shape, traits and behaviour: its phenotype. The two are worlds apart: it’s rather like trying to understand the meaning of a Dickens novel from a close consideration of the shapes of its letters and the correlations in their order of appearance.
Besides, this conventional “blueprint” description of what genomes do is too simplistic even if we consider only how they dictate that roster of proteins. Here are some reasons why:
Only about 1.5 per cent of the human genome encodes proteins, and a further 8 to 15 per cent or so is thought to “regulate” the activity of other genes by encoding RNA that turns their transcription up or down. We don’t know what the rest does, and scientists aren’t agreed on whether it is just useless “junk” accumulated, like rubbish in the attic, over the course of evolution, or whether it has some unknown but important biological function. In all probability, it is a bit of both. But at any rate, a lot of this DNA with no known protein-coding or regulatory function is nonetheless transcribed by cells to RNA, and no one is sure why.
Most protein-coding human genes each encode more than one protein. Genes are not generally simply a linear encoding of protein sequences that start at one end of the protein chain and finish at the other; they are, for example, interspersed with sequences called introns that are carefully snipped out of the transcribed RNA before it is translated. Sometimes the transcribed RNA then gets reshuffled before translation, providing templates for several different proteins.
Proteins are not just folded chains of amino acids. Sometimes those folded chains are “stapled” in place by chemical bonds, or clipped together by other chemical entities such as electrically charged ions. Most proteins have other chemical groups added to them (by other enzymes) – for example, a group containing an iron atom is needed by the protein haemoglobin to bind oxygen and carry it around the body in the blood. None of these details, essential to the protein’s structure and function, is encoded in DNA. You would not be able to deduce them from a gene sequence.
We only know what around 50 per cent of gene-encoded proteins do, or even what they look like. The rest are sometimes called “dark” proteins: we assume they have a role but we don’t know what it is.
Plenty of proteins do not seem to have well-defined folded states but appear loose and floppy. Understanding how such ill-defined “intrinsically disordered proteins” can have specific biological roles is a very active area of current research. Some researchers think that the floppiness may not reflect the state of these proteins in cells themselves – but we don’t really know if that is so or not.
Ah, details, details! How much should we care? Do they really alter the picture of genes dictating the organism?
That depends, to some degree, on what questions you are asking. A genome sequence – the ordered list of nucleotides A, T, C and G along the DNA strands of chromosomes – does specify the nature of the organism in question. From this sequence you can tell in principle if the cell that contains it is from a human, a dog or a mouse (something that may not be obvious from a cursory look at the cell as a whole). These distinctions are found only in some key genes: the human genome differs from that of chimpanzees in just 1 per cent of the sequence, and a third of it is essentially the same as the genome of a mushroom.8 The differences between the genomes of individual people are even tinier.
But whereas you can look at a real blueprint, and probably an instruction manual, and figure out what kind of object will emerge from the plan, you can’t do that for a genome. Indeed, you can only deduce that a genome will “produce” a dog at all if you have already decoded the generic dog genome for comparison, laying the two side by side. It’s simply a case of seeing if the two genomes superimpose; there’s nothing intrinsic in the sequence that hints at its “dogness”.
This isn’t because we don’t yet know enough about the “instructions” in a genome (although that is the case too). It’s because there is no direct relationship between the informational content of a gene – which, as I say, typically dictates the structure of a class of protein molecules, or at least of the basic amino-acid fabric of those proteins – and a trait or structure apparent in the organism. Most proteins do jobs that can’t easily be related to any particular trait. Some can be: for example, there’s a protein that helps chloride ions get through the membranes of our cells, and if this protein is faulty – because of a mutation in the corresponding gene – then the lack of chloride transport into cells causes the disease cystic fibrosis. But in general, proteins carry out “low-level” biochemical functions that might be involved in a whole host of traits, and which might have very different outcomes if the protein is produced (“expressed”) at different stages in the development or life cycle of the organism. As microbiologist Franklin Harold has said, “the higher levels of order, form and function are not spelled out in the genome.”
Might we, then, call a genome not a blueprint but a recipe? The metaphor has rather more appeal, not least because many recipes assume implicit knowledge (especially in older cookbooks). But a recipe is still a list of ingredients plus instructions to assemble them. Genomes do not come with users’ instructions, more’s the pity. Harold offers a different image, allusive and poetic and all the more appealing for that:
I prefer to think of the genome as akin to Hermann Hesse’s Magister Ludi [aka The Glass Bead Game]: master of an intricate game of cues and responses, in which he is fully enmeshed and absorbed; a game that is shaped as much by its own internal rules as by the will of that masterful player.
If there was better public communication of the complex, contingent and often opaque relationship of genotype to phenotype, there might be rather less anxiety about the idea that genes affect behaviour. Small variations in each individual’s genetic make-up can have an influence – sometimes a rather strong one – not just what you look like but what your behaviour and personality are like. This much is absolutely clear: there is not a single known aspect of human behaviour so far investigated that does not turn out to show some correlation