In 1994 Daniel Nilsson and Susanne Pelger used a computer to see what would happen to a mathematical model of a light-sensing surface if it was allowed to change in small, random, biologically feasible ways, with only those changes that improved its sensitivity to light being retained. They found that within 400,000 generations - an evolutionary blink of an eye - that flat surface gradually changed into a recognizable eye, complete with a lens. The lens even bent light differently in different places, just like our eye and unlike normal spectacle lenses. At every tiny step along the way, a creature with the improved ’eye’ would be better than those with the old version.
At no stage was there ever ’half an eye’. There were just light-sensing things that got better at it.
Since the 1950s, we have been in possession of a new and central piece of the evolutionary jigsaw, one that Darwin would have given his right arm to know about. This is the physical - more precisely, chemical - nature of whatever it is that ensures that characteristics of organisms can change and be passed from one generation to the next.
You know the word: gene.
You know the molecule: DNA.
You even know how it works: DNA carries the genetic code, a kind of chemical ’blueprint’ for an organism.
And, probably, a lot of what you know is lies-to-children.
Just as ’survival of the fittest’ captured the imaginations of the Victorians, so ’DNA’ has captured the imaginations of today’s public. However, imaginations thrive best if they are left free to roam: they grow tired and feeble in captivity. Captive imaginations do breed quite effectively, because they are protected from the terrible predator known as Thought.
DNA has two striking properties, which play a significant role in the complex chemistry of life: it can encode information, and that information can be copied. (Other molecules process the DNA information, for example by making proteins according to recipes encoded in DNA.) From this point of view a living organism is a kind of molecular computer. Of course there’s much more to life than that, but DNA is central to any discussion of life on Earth. DNA is life’s most important molecular-level ’space elevator’ - a platform from which life can launch itself into higher realms.
The complexity of living creatures arises not because they are made from some special kind of matter- the now-discredited ’vital-ist’ theory but because their matter is organized in an exceedingly intricate fashion. DNA does a lot of the routine ’bookkeeping’ that keeps living creatures organized. Every cell of (nearly) every living organism contains its ’genome’ - a kind of code message written in DNA, which gives that organism a lot of hints about how to behave at the molecular level. (Exceptions are various viruses, on the boundary between life and non-life, which use a slightly different code.)
This is why it was possible to clone Dolly the Sheep - to take an ordinary cell from an adult sheep and make it grow into another sheep. The trick actually requires three adult sheep. First, there’s the one from which you take the cell: call her ’Dolly’s Mum’. Then you persuade the cell’s nucleus to forget that it came from an adult and to think that it’s back in the egg, and then you implant it into an egg from a second sheep (’Egg Donor’). Then you put the egg into the uterus of the third sheep (’Surrogate Mum’) so that it can grow into a normal lamb.
Dolly is often said to be a perfect copy of Dolly’s Mum, but that’s not completely true. For a start, certain parts of Dolly’s DNA come not from Dolly’s Mum, but from Egg Donor. And even if that slight difference had been fixed, Dolly could still differ in many ways from her ’mother’, because sheep DNA is not a complete list of instructions for ’how to build a sheep’. DNA is more like a recipe - and it assumes you already know how to set up your kitchen. So the recipe doesn’t say ’put the mixture in a greased pan and place in an oven set to 400°F,’ for instance: it says ’put the mixture in the oven’ and assumes that you know it needs to go in a pan and that the oven should be set to a standard temperature. In particular, sheep DNA leaves out the vital instruction ’put the mixture inside a sheep’, but that’s the only place (as yet) where you can turn a fertilized sheep egg into a lamb. So even Surrogate Mum played a considerable role in determining what happened when the DNA recipe for Dolly was ’obeyed’.
Many biologists think that this is a minor objection after all, Egg Donor and Surrogate Mum work the way they do because their DNA contains the information that makes them do it. But things that aren’t in any organism’s DNA may be essential for the reproductive cycle. A good example occurs in yeast, a plant that can turn sugar into alcohol and give off carbon dioxide. The entire DNA code for one species of yeast is now known. Thousands of experimentalists have played genetic games with yeast, then spun the beasties in a centrifuge to separate the DNA, from which they can work out the code. When you do this, you leave a scummy residue in the bottom of the test tube, but since it’s not DNA, you know it can’t be important for genetics, and you throw it away. And so they all did, until in 1997 one geneticist asked a stupid question. If it’s not DNA, what’s it for? What’s in that scummy residue, anyway?
The answer was simple, and baffling. Prions. Lots and lots of them.
A prion is a smallish protein molecule that can act as a catalyst for the formation of more protein molecules just like itself. Unlike DNA, it doesn’t do this by replication. Instead, it needs a supply of proteins that are almost like itself, but not quite the right atoms, in the right order, but folded into the wrong shape. The prion attaches itself to such a protein, jiggles it around a bit, and nudges it into the same shape as the prion. So now you’ve got more prions, and the process speeds up.
Prions are molecular preachers: they make more of themselves by converting the heathen, not by splitting into identical twins. The most notorious prion is the one that is believed to be the cause of BSE, ’mad cow disease’. The protein that gets converted happens to be a key component of the cow’s brain, which is why infected cows lose coordination, stagger around, foam at the mouth, and look crazy. What does yeast want prions for? Without prions, yeast can’t reproduce. The protein-making instructions in its DNA sometimes make a protein that is folded into the wrong shape. When a yeast cell divides, it copies its DNA to each half, but it shares the prions (which can be topped up by converting other proteins). So here’s a case where, even on the molecular level, an organism’s DNA does not specify everything about that organism. There’s a lot about the DNA code system that we don’t understand, but one part that we do is the ’genetic code’. Some segments of DNA are recipes for proteins. In fact, they come very close to being exact blueprints for proteins, because they list the precise components of the protein and they list them in exactly the right order. Proteins are made from a catalogue of fairly tiny molecules known as amino acids. For most organisms, humans included, the catalogue contains exactly 22 amino acids. If you string lots of amino acids together in a row, and let them fold up into a relatively compact tangle, you get a protein. The one thing the DNA doesn’t list is how to fold the resulting molecule up, but usually it folds the right way of its own accord. Occasionally, when it doesn’t, there are more servant molecules to nudge it into the right shape. Just such a servant molecule, rejoicing in the name HSP90, is turning molecular genetics upside down even as we write. HSP90 ’insists’ that proteins fold into the orthodox shape, even if there are a few mutations in the DNA that codes for those proteins. When the organism is ’stressed’, diverting HSP90 to other functions, these cryptic mutations suddenly get expressed - the proteins acquire the unorthodox shape that goes along with their mutated DNA codes. In effect, this says that you can trigger a genetic change by non-genetic means.
Segments of DNA that code for working proteins are called genes. Segments that don’t rejoice in a variety of names. Some of them code for proteins that control when a given gene ’switches on’, that is, starts to make proteins: these are known as regulatory (or homeotic) genes. Some bits are colloquially called ’j
unk DNA’, a scientific term meaning ’we don’t know what these bits are for’. Some literally minded scientists read this as ’they’re not for anything’, thereby getting the horse of nature neatly aligned with the rear end of the cart of human understanding. Most likely they are a mix of different things: DNA that used to have some function way back in evolution but currently does not (and might possibly be revived if, say, an ancient parasite reappeared), DNA that controls how genes switch their protein manufacturing on and off, DNA that controls those, and so on. Some may actually be genuine junk. And some (so the joke goes) may encode a message like ’It was me, I’m God, I existed all along, ha ha.’
Evolutionary processes do not always direct themselves along paths that are neatly comprehensible to humans. This doesn’t mean Darwin was wrong: it means that even when he’s right, there may be a surprising absence of narrativium, so that a ’story’ that makes perfect sense to evolution may not make sense to humans. We suspect that a lot of what you find in living organisms is like that -offering a small advantage at every stage of its evolution, but an advantage in such a complex game is that we can’t tell a convincing story about why it’s an advantage. To show just how bizarre evolutionary processes can be, even in comparatively simple circumstances, we must look not to animals or plants, but to electronic circuits.
Since 1993 an engineer named Adrian Thompson has been evolving circuits. The basic technique, known as ’genetic algorithms’, is quite widely used in computer science. An algorithm is a specific program, or recipe, to solve a given problem. One way to find algorithms for really tough problems is to ’cross-breed’ them and apply natural selection. By ’cross breed’ we mean ’mix parts of one algorithm with parts of the other’. Biologists call this ’recombination’ and each sexual organism - like you - recombines its parents’ chromosomes in just this manner. Such a technique, or its result, is called a genetic algorithm. When the method works, it works brilliantly; its main disadvantage is that you can’t always give a sensible explanation of how the resulting algorithm accomplishes whatever it does. More of that in a moment: first we must discuss the electronics.
Thompson wondered what would happen if you used the genetic algorithm approach on an electronic circuit. Decide on some task, randomly cross-breed circuits that might or might not solve it, keep the ones that do better than the rest, and repeat for as many generations as it takes.
Most electronic engineers, thinking about such a project, will quickly realize that it’s silly to use genuine circuits. Instead, you can simulate the circuits on a computer (since you know exactly how a circuit behaves) and do the whole job more quickly and more cheaply in simulation. Thompson mistrusted this line of argument, though: maybe real circuits ’knew’ something that a simulation would miss.
He decided on a task: to distinguish between two input signals of different frequencies, 1 kilohertz and 10 kilohertz - that is, signals that made 1000 vibrations per second and 10,000 vibrations per second. Think of them as sound: a low tone and a high tone. The circuit should accept the tone as input signal, process it in some manner to be determined by its eventual structure, and produce an output signal. For the high tone, the circuit should output a steady zero volts - that is, no output at all - and for the low tone, the circuit should output a steady five volts. (Actually, these properties were not specified at the start: any two different steady signals would have been acceptable. But that’s how it ended up.)
It would take forever to build thousands of trial circuits by hand, so he employed a ’field-programmable gate array’. This is a microchip that contains a number of very tiny transistorized ’logic cells’ - mildly intelligent switches, so to speak - whose connections can be changed by loading new instructions into the chip’s configuration memory.
Those instructions are analogous to an organism’s DNA code, and can be cross-bred. That’s what Thompson did. He started with an array of one hundred logic cells, and used a computer to randomly generate a population of fifty instruction codes. The computer loaded each set into the array, fed in the two tones, looked at the outputs, and tried to find some feature that might help in evolving a decent circuit. To begin with, that feature was anything that didn’t look totally random. The ’fittest’ individual in the first generation produced a steady five-volt output no matter which tone it heard. The least fit instruction codes were then killed off (deleted), the fit ones were bred (copied and recombined), and the process was repeated.
What’s most interesting about the experiment is not the details, but how the system homed in on a solution - and the remarkable nature of that solution. By the 220th generation, the fittest circuit produced outputs that were pretty much the same as the inputs, two waveforms of different frequencies. The same effect could have been obtained with no circuit at all, just a bare wire! The desired steady output signals were not yet in prospect.
By the 650th generation, the output for the low tone was steady, but the high tone still produced a variable output signal. It took until generation 2800 for the circuit to give approximately steady, and different, signals for the two tones; only by generation 4100 did the odd glitch get ironed out, after which point little further evolution occurred.
The strangest thing about the eventual solution was its structure. No human engineer would ever have invented it. Indeed no human engineer would have been able to find a solution with a mere 100 logic cells. The human engineer’s solution, though, would have been comprehensible - we would be able to tell a convincing ’story’ about why it worked. For example, it would include a ’clock’ - a circuit that ticks at a constant rate. That would give a baseline to compare the other frequencies against. But you can’t make a clock with 100 logic cells. The evolutionary solution didn’t bother with a clock. Instead, it routed the input signal through a complicated series of loops. These presumably generated time-delayed and otherwise processed versions of the signals, which eventually were combined to produce the steady outputs. Presumably. Thompson described how it functioned like this: ’Really, I don’t have the faintest idea how it works.’
Amazingly, further study of the final solution showed that only 32 of its 100 logic cells were actually needed. The rest could be removed from the circuit without affecting its behaviour. At first it looked as if five other logic cells could be removed - they were not connected electrically to the rest, nor to the input or output. However, if these were removed, the circuit ceased to work. Presumably these cells reacted to physical properties of the rest of the circuit other than electrical current - magnetic fields, say. Whatever the reason, Thompson’s hunch that a real silicon circuit would have more tricks up its sleeve than a computer simulation turned out to be absolutely right.
The technological justification for Thompson’s work is the possibility of evolving highly efficient circuits. But the message for basic evolutionary theory is also important. In effect, it tells us that evolution has no need for narrativium. An evolved solution may ’work’ without it being at all clear how it does whatever it does. It may not follow any ’design principle’ that makes sense to human beings. Instead, it can follow the emergent logic of Ant Country, which can’t be captured in a simple story.
Of course, evolution may sometimes hit on ’designed’ solutions, as happens for the eye. Sometimes it hits on solutions that do have a narrative, but we fail to appreciate the story. Stick insects look like sticks, and their eggs look like seeds. There is a kind of Discworld logic to this, since seeds are the ’eggs’ of sticks, and prior to the theory of evolution taking hold the Victorians approved of this ’logic’ because it looked like God being consistent. The early evolutionists didn’t see it that way, and they worried about it; but they worried a lot more when they found that some stick insect eggs looked like little snails. It seemed silly for anything to resemble the favourite food of nearly everything else. In fact, it seemed to be a flat contradiction to the evolutionary story. The puzzle was solved only in 1994, af
ter forest fires in Australia. When new plant shoots came up out of the ashes, they were covered in baby stick insects. Ants had carried the ’seeds’, and the ’baby snails’, down into their subterranean nests, thinking they were the real thing. Being safely underground, the stick insect eggs escaped the fires. In fact, baby stick insects look, and run, just like ants: this should have been a clue, but nobody made the connection.
And sometimes evolution’s solution has no narrative structure. To test Darwin’s theories thoroughly, we should be looking for evolved systems that don’t conform to a simple narrative description, as well as for ones that do. Many of the brain’s sensory systems may well be like this. The first few layers of the visual cortex, for example, perform generalized functions like detecting edges, but we have no idea how lower layers work, and that may well be because they don’t conform to any design principles that we currently can recognize. Our sense of smell seems to be ’organized’ along very strange lines, not at all as clearly structured as the visual cortex, and it too may be lacking any element of design.
More importantly, genes may well be like this. Biologists habitually talk of ’the function of a gene’ what it does. The unspoken assumption is that it does only one thing, or a small list of things. This is pure magic: the gene as a spell. It is conceived as being a spell in the same sense that ’Cold Start’ in a car is. But a lot of genes may not do anything that can be summed up in a simple story. The job they evolved to do is ’build an organism’, and they evolved as a team, like Thompson’s circuits. When evolution turns up solutions of this kind, conventional reductionism is not much help in understanding those solutions. You can list neural connections till the cows come home, but you won’t understand how the cows’ visual systems distinguish a cowshed from a bull.
TWENTY-SEVEN
WE NEED MORE BLOBS
Terry Pratchett - The Science of Discworld Page 22