(electric music) (audience applauding) - Thank you. Thank you, it is so nice to be back in the finest science
lecture theatre in the world. And I've been giving talks
here for over 20 years now, I realise, but I don't think I've ever had a crowd quite like this. So I feel very flattered and
a little bit nervous tonight. There are times, aren't
there, when we could really do with an instruction manual for the human. Just imagine if whenever
something goes wrong, we could just look up the
correct fix in the manual. A viral infection, an allergy, whatever, we just look it up. But the truth is that after
pondering this question for thousands of years, we still only got a very
sketchy idea of the answer. Now, you might think that
to answer this question, we should really start with another one. And some very smart people
have pondered that question over the years, but still, no one agrees on the answer. I'm going to go, for
the time being at least, with the answer that the British biologist J.B.S. Haldane gave at the
start of his 1947 essay with this title where he simply said this. (audience laughs) So we can park that question for now, I'll come back to it, but the fact is that I know I'm alive and I know you're alive, and so we can ask how do we work? Well, this is where we'll
start with that question, down the pub, specifically down the
Eagle Pub in Cambridge which was where, according to the American biologist James Watson, he and Francis Crick burst in in 1953 for Crick to exclaim to everyone, "We have discovered the secret of life." Watson recently admitted that actually he just made up this story for his book. Despite having unveiled
this plaque to that effect, he made it up
(audience laughs) for dramatic effect in his
1968 book, "The Double Helix". For it was that double helix
molecule that he and Crick had just deduced the
structure for the DNA, the double helix of the DNA molecule. And Watson wanted us to believe clearly that this was the secret
of how life works. DNA was then widely
thought, not by everyone, but very widely thought to be
where the genetic information that is passed between
generations is stored. And in this double helical structure, Crick and Watson had seen how it can be that our DNA is able to store
this genetic information. Each of the twin strands
of the DNA double helix is a string of a succession of just four different chemical building blocks, which we denote by the initial
letters of their names. So C, G, T, and A. And this, they said,
then acts like a code, just like the strings of binary digits, the ones and zeros that
were then being used to encode information on the magnetic tape that was used by digital computers. So, each of our genes then is a short segment of this code, and we have many thousands of these genes in the string of letters,
chemical letters, that for we, humans,
stretches to 3 billion of these letters long in
which we call the genome. So here, then, was apparently how we work. The single fertilised egg
from which we all came comes loaded with a set of
instructions in the genome and all that needs then to happen is that these instructions are
read out to build us. So all we needed to do now
was to read the instructions. And that was what was done
in the Human Genome Project which began in 1990 and
was completed at least as a rough draught by the
turn of the millennium. As Bill Clinton said when that draught was first publicly announced, "Today, we are learning the language in which God created life." So here it is, our instruction
booklet as it's often called, or our blueprint, or the letters of the human genome, which you can go and read,
should you be inclined, in these 109 volumes held in the Wellcome Collection on Euston Road. Is this then the secret of how life works? Well, some people have questioned whether this blueprint or
instruction booklet metaphor is really the right way
to think about the genome. The Oxford physiologist
Denis Noble has suggested that perhaps a better analogy
might be a musical score. And that's certainly better,
I think in some respects. (orchestral music) So what's the cause of this music? Well, some might say it's the
score composed by Beethoven But is it? That answer reminds me of the comment by the Russian violinist Jascha Heifetz when he was once
complimented by an admirer after a concert he gave, who came up and said, "Your violin has such a beautiful tone." And Heifetz picked up
his instrument and said, "I don't hear anything." (audience laughs)
And his point of course was that that beautiful tone doesn't just sort of happen. It's not inherent in the violin. It takes something more to produce it. By the same token, if
you put the human genome, if you just put it in a glass of water it'll just sit there. It'll never make life at all. It won't make a single cell, let alone us. Well, obviously, right? I mean, obviously you need an orchestra to actually get any music. And so one reason why
Denis's metaphor, I think, is quite nice is that it helps one see it's not so much that the musical score is the cause of the music, but that different musical scores, when played by an orchestra, account in some sense for the differences between the music they will make. So the problem with saying that the genome is the secret of life is
evidently that the genome is not what makes life happen. Sure, life won't happen without it, just like Beethoven's "3rd
Symphony" wouldn't happen unless Beethoven had prepared a score. But what we're really saying
is that for life to happen, the genome needs to go into
an already living system. So it's really, the story is,
(orchestral music) oops, sorry, I move it
quickly on from there. Didn't know that would... Oh, it was hovering over the, (chuckles) over the music. So this is really the right
way to tell the story. Life is, in other words,
a genome plus life. (audience laughs)
Or to put it another way, this is the picture
that we're often given. And we're told what goes
on in this black box is so horrendously contemplated that you're probably better off not looking inside it at all. But just rest assured that
scientists are working on it, and one day they'll
have it all sorted out. However, that advice neglects
this inconvenient fact. So, to understand how life works, I'm afraid that, yes, we really do have to take a peek inside this black box. And I'm going to try to give you a little glimpse of it inside tonight. And you know, I won't deny
that in all its glory, it really is horrendously complicated. But I hope all the same to persuade you that we can make some sense of it. And what's more, that
over the two decades or so since the Human Genome
Project was completed, the sense we've made tells
a rather different picture from the one that we have
traditionally been told in the genomic era. Well, Francis Crick opened up
that black box a little bit, and this is what he saw, This is what, in the late 1950s, he called this central
dogma of molecular biology. Now, that's an odd name, isn't it? Because science isn't
meant to have any dogmas. It's meant to be provisional
and subjective revision, right? But Crick later admitted
that he called it that because he didn't really
know what dogma meant. (audience laughs)
But all the same, calling something a dogma in science is a bit like a red flag to a bull because it has goaded a lot of people to devote a lot of time and energy to trying to argue and to show that Crick's central dogma is wrong. I'd say that a better
way to look at it is that it's not so much that it's wrong, but it's a bit like the Holy Roman Empire, which was once famously
said to be not holy, not Roman, and not really an empire. So the central dogma is not so much wrong as that it's not a dogma and it's not really so central. But let's take a look at it. It sums up what was already
thought to be the case and what was soon proved to be the case that what the genes really encode is instructions for making
these molecules called proteins. And you probably remember the
usual cliche about proteins. They are the workhorses of the cell, which means that they're the molecules that make the cell's
biochemical reactions happen generally by acting as catalysts that help convert one biological
molecule into another. And the way the code in
the genes is read out and turned into proteins
is a two-step process. So the first step is
that the code in a gene, a given gene, is copied, and the technical word is transcribed, into a very similar
molecule to DNA called RNA. So just that little bit of the genome is transcribed into this
so-called messenger, RNA molecule, and then this mRNA moves off and is seized by another
piece of molecular machinery in the cell called the ribosome, which uses the information
that the RNA encodes to put together a particular
string of amino acids in a so-called polypeptide chain, which then folds up
into this compact shape, and that's the protein molecule. So the protein has a particular shape, and this is the process
called translation. And there are thousands and
thousands of these proteins in every one of our cells. Here's a snapshot of the
inside of one of those cells. And this isn't just a
sort of computer cartoon of what it might look like. This is a computer generated image using actual data from inside a cell. So this is what it's like
inside every cell of our bodies and you can see it is insanely crowded. But no matter, the usual story goes because each of those
proteins has, as I say, a very particular shape
which fits together like a lock and key with the molecule that it's meant to transform. So, it will only go to
work on that molecule and it will ignore all the others. And through this series of highly specific molecular interactions, proteins somehow put together us. Well, here's how that will be expressed within the context of
Crick's central dogma. So this fancy word here, phenotype, basically it just means
all the traits we have. What forms we take, how tall we are, what colour our skin and our eyes are, even what behaviours we have. The phenotype is basically
our individual selves. And in this picture, it all
comes from what's encoded in the genes, in the so-called genotype. But since the Human Genome
Project was completed, more and more complications and problems with this story have emerged, and I want to tell you about some of them. So I said that we have thousands of genes and they make thousands
of different proteins. How many genes exactly do we have? Well, a banana has 36,000 genes. Now, think very carefully
before you answer this question. Which organism is the more complicated? (audience laughs) Okay, so it stands to reason that I'm going to need more
than 36,000 genes, right? When the Genome Project started, a common figure that scientists would give for the estimated number of human genes was around 80 to a 100,000. So I've shown it in a dotted line here 'cause it's just an estimate. But we pretty soon got
a sobering wake up call, because as we started
getting into the project those numbers fell dramatically. And when we actually
started getting real data, the solid line here, we found that it is much lower, that it was something like 20,000 genes that encode proteins. And now, some scientists think it might be even slightly lower, might
be as many as 19,000. So that's about as many as a tiny soil dwelling worm
called the nematode has, and it's scarcely half what a banana has. And these numbers are often
paraded now as a comic example of how wrong expert opinion can be. But I think the really
important question is why they were wrong? Did we perhaps have the wrong idea about what role genes were playing? Well here's another thing. In the 1990s, there were one or two genes for which biologists didn't seem able to find the corresponding protein. And in the end, they had to conclude that that's because there aren't any. These genes simply make RNA, and not the messenger
RNA that gets translated by the ribosome into a protein. The RNA itself is the end. It has some biochemical function, it does the kind of things
that we thought proteins do. And these RNA encoding genes, and I've just marked them in red here, they're called non-coding genes. Not clearly because they
don't encode anything, but because their genes
don't encode proteins, which is what we thought all genes did. Well, okay, but biology is full
of weird exceptions, right? Except that these weren't exceptions. Over the past two decades, the number of these non-coding
genes kept creeping up and eventually, just a few years back, that number exceeded the
number of protein-coding genes. And what's more, current estimates are that it's just going to continue that way. That actually it's going to turn out that there are lots more
of these non-coding genes, that they vastly outnumber
the protein-coding genes. And the picture is actually
even more transformed than that because these are just
the pieces of the genome that encode these relatively
long, non-coding RNA molecules that qualify as genes. But it's been discovered
that there are actually lots of other bits of the
genome, or our genome, but also those of other
large animals like us, called metazoans, that encode
lots of smaller RNA molecules. And there are all these different families with these fancy names
that do all sorts of tasks in the cell, protein-like tasks. So, you know, the genome is full of things that don't encode proteins. So you can see, the genome isn't really what
we thought it was about, and yet somehow we still
seem to be telling, certainly hearing, the same story about it that was being told back in
the 1990s or even the 1970s. That can't be right. It's rather as if
cosmologists were to have said after discovering that
four-fifths of the matter in the universe is made up
of this so-called dark matter that we can't see and about
which we know nothing, as if they just sort of shrugged and said, "This change is nothing." Fortunately they didn't say that. Well, it's actually even worse than this. Crick's central dogma said
that genes encode proteins, and by that we mean
that the genes actually programme the proteins
with particular shapes so that they can go and
do their specific jobs. But, here's one way in which
this picture is now modified. It doesn't mean that each gene
encodes a particular protein. In fact, each of our genes
can typically be used to make several different proteins. On average, each can make
about six different proteins, but some genes can encode dozens or even hundreds of different proteins. So we have many more proteins and no one knows exactly how many more, but many more than we
have protein-coding genes. How is that possible? Well, it's because, as was
first discovered in the 1970s, the messenger RNA that is
transcribed from a gene is typically chopped up and edited before it is translated. So there's another piece
of molecular machinery, this thing called the spliceosome, made up of several different proteins that gets hold of the messenger RNA, chopped it into fragments, throws away some pieces called introns, and stitches together
the remaining fragments called exons back in various orders. What decides how this
editing and splicing occurs is typically information coming from a higher level of the system. Say, for example, from the
overall state of the cell in which it's happening. So, a gene in one tissue might
produce one type of protein and in a different tissue might
produce a different protein. In other words, the information flow here isn't, as the central
dogma at least implied, isn't all from the bottom up
from DNA to RNA to proteins. Some crucial information
for making the proteins is coming from the outside in some sense. And this is just one of the ways in which in order to build
us and to keep us alive all these years, information doesn't just flow upwards from the genes to higher
levels of organisation, but flows up and down and in between and in all sorts of directions among them. It's an open informational system, not a closed one. And here's another change to the picture. In the analogy of the
genome to a musical score, we might say that the score is what prevents the
orchestra from just playing a whole load of random notes, right? Making a racket because
it tells each musician exactly which notes to play and when. And that's the equivalent of the way a protein's gene encoded shape, like this one, tells it
what to do in the cell, which molecules to grab
hold of and which to ignore. So I mentioned earlier
this lock and key aspect. But we now know that for
many of our proteins, including some with some of the most important jobs in the cell, the DNA score isn't like this at all. It's much more open to interpretation. And what I mean by that is that some genes encode proteins without
assigning them a structure. It leaves them loose and floppy, or as biochemists say they
are intrinsically disordered. And this isn't some failure of the genome to give proteins a proper shape. It's clearly a deliberate feature that evolution has, so to speak, chosen. Because you see there's much
less of this intrinsic disorder among the proteins of simpler
organisms like bacteria. So evolution clearly was fine giving proteins all of a
very specific structure, but it seems that it has found it useful or perhaps even necessary
to give proteins disorder in order to make more
complex, multicellular, multi-tissue organisms like us. Given that so many of our proteins, but perhaps a third to a half or maybe even more of them, have parts or holes that
are disordered in this way, given that, it seems a little odd that we didn't really know about this until the past few decades. But that's because the methods that scientists have in the past used to look at protein structure generally only work well
for the kinds of proteins that are ordered, that
have a fixed structure. And so they'll pack together
and form nice orderly crystals which are what you need for those methods. It's a bit like the way we overlooked non-coding genes for so long. We tend to see only the
things that we expect to see and we tend to study only those things that we
have techniques for studying. So proteins with this intrinsic disorder, these floppy proteins, are far less choosy about which other molecules they stick to. They're rather indiscriminately sticky. That's to say, they're quite promiscuous in their molecular unions. So, that old idea that the
molecular chaos of the cell is somehow kept orderly and tamed because each of the
proteins is highly selective about what it interacts with, that idea doesn't really work. And let me show you a
particularly important example of this kind of molecular
promiscuity in action. So, the standard argument for
how we get to be so complex with so few protein-coding
genes goes something like this. That we say, well, it
arises from the complexity of all the different interactions
between those molecules that interact in these really
horrible looking networks, where, you know, each of these
blobs represents a protein. And it looks pretty horrid, doesn't it? If you open a copy of Nature at random, you're bound to see
pictures kind of like this, and it's tempting to
think of them as cartoons of what all the various
molecules are actually doing, moving around the cell. But once you remember how complicated a cell really is inside, you have to wonder how on earth a complex dance of molecules like this could be orchestrated. All the same, the idea is
that this molecular crosstalk explains how it is that, for example, different genes are turned on and off in different types of cell in our bodies. Making heart muscle cells different from skin cells or liver cells even though they all have the same genome. The idea is that proteins made by one gene could, for example,
act as a kind of switch to control the transcription
of another gene, so that it turns on or
off whether that gene is transcribed and translated. And that process is
called gene regulation. Now, we've known for a long
time, since at least 1960s, that gene regulation happened. It was around that time that the French biochemists
Jacques Monod and Francois Jacob showed how this process worked for one particular type of gene regulation in the bacterium E. coli. Now, E. coli can digest two
different types of sugar. It can digest glucose and lactose, but it's not very
efficient if the bacterium is constantly making both
of the two enzymes needed for those two processes when there's only one
sugar or the other around. And so there's a switch, which Monod and Jacob
called the lac operon, a switch for turning on and off the production of the
so-called Lac enzymes that are the ones that digest lactose. So in short, this is what happens. There's this protein,
this green thing here, that can recognise and stick to a little patch on the
DNA, this yellow patch, just before the lac genes themselves. And if it sticks there, then it blocks this pink
blob, the RNA polymerase, that produces transcription
that produces RNA. It blocks it and sort of kicks it off so that it can't do its job. And so it stops the
transcription of the lac genes. So, it's like a nice,
simple digital switch. There's a transparent logic to it. And for a long time,
molecular biologists figured that gene regulation in organisms like us follows the same kind of principles, effectively wiring our
genes into a network a bit like a digital circuit like we have in microelectronic devices. Well, you can probably
guess what I'm gonna say. That's not how it turned out. (audience laughs)
Very often, gene regulation in metazoans like us is much more complicated. These proteins that
interact with parts of DNA to control gene expression are called transcription factors. And it turns out that many
of our transcription factors are intrinsically disordered proteins, which aren't so selective
as these bacterial proteins in what they bind to, maybe in which bits of DNA they bind to or which are the molecules. And the bits of DNA
that regulate our genes, like this sort of yellow section here, there are lots of them in our genes and they're not all next
to the genes they regulate. Some of them, weirdly, are a long way away on the DNA strand, somewhere else entirely. And these are regions called enhancers that somehow control the extent to which the gene is switched on or off. So gene regulation in us tends to involve a whole bunch of different components, transcription factors and other molecules, often including some of those
non-coding RNA molecules. Most of them are regulatory. They have functions in gene regulation. And then there are other things. There are all these bits of DNA that are controlling the process somehow, some of which because they are so far away they are brought close by pulling out these big loops of DNA and sort of looping them back around, like tugging out a piece of
wool from a tangled ball. And all of these components fit together into this kind of gigantic
regulatory assembly that isn't something that just kind of clips together neatly. It's a loose disorderly blob, a dense cluster sometimes
called a condensate, which forms like a kind of liquid droplet, a bit like a blob of vinegar
in the oil of salad dressing. So, you know, it looks
like a really messy way to do this job. It's as though gene regulation in us is done by these loose
committees of molecules all talking to each other
rather indiscriminately. And it really, it's pretty
amazing that somehow all of these components still manage to make a reliable decision about whether to switch
gene expression up or down despite all this fuzziness
amongst their interactions. This sort of fuzziness and the way it produces
collective decisions rather than simple, digital
logic that we see in bacteria based on precise molecular unions, this fuzziness is a characteristic feature of our molecular biology. We don't quite understand how it works, but I think we can start to see why it is that for us life does work in this fuzzy analogue and rather open-ended way. You see, once you start to think about it, the more complex an organism is, the more a blueprint or an
instruction book approach to controlling how it works starts to look like a terrible idea. If the organism working correctly depending on each of those instructions being executed perfectly
at just the right place and just the right time in some sort of complex clockwork manner, then it's just not going to happen. Not least because the molecular world isn't like clockwork machinery, it's full of randomness and noise. It would be like trying to, expecting a mechanism like this to go on working perfectly if you were to immerse it in
the bath and shake it around. It's not going to happen. I mean, sure, you can, and
this was often the story that was told before, you can build incontingency plans. So if one bit fails, there's another way for the same thing to happen. So the idea was that in
these complex networks, there's more than one
route to the same end, so that if this route fails,
there's always this one. But when you think about it, that's not a great way
to solve the problem. The answer to the fragility that comes from something that's very, very complex If it has to all work perfectly, the answer to that problem can't be to just give it more complexity. Instead, you need to use totally
different design principles for making it. And that's surely why we have these fuzzy molecular mechanisms, where the details often
don't really matter. The committees can still
come to good decisions even if some of the members
are absent or asleep. What these principles really amount to, what this idea really amounts to is taking the responsibility for the correct functioning
of the whole thing off the lower levels of the system and handing it up to some higher level. In other words, we work
in a way that is designed to take the pressure off our genes. To make them, in general,
no longer the real cause of our traits and
behaviours of our phenotype. So, mistakes and malfunctions
at the lower levels can be compensated for higher up. And this happens actually
at other sort of levels in the stratum in the hierarchy of the way complex organisms like us work. So that, for example, if cells during the development of an embryo are sort of don't quite end
up where they're supposed to, often there's a way to compensate for that further down the developmental line so that you still end up with
a perfectly viable organism. And let me give you briefly
a couple of examples of how these higher level principles that help organise our tissues and bodies reveal this kind of
dispersal of responsibility so that they involve genes without in any sense
being blueprinted by them. So, the surfaces of our
intestines are covered with these little finger-like
protrusions called villi, which hugely increase its surface area so that it can absorb
nutrients sufficiently into the bloodstream. And these are basically bulges in the surface layer of
tissue called the epithelium, and their growth is triggered by a protein that rejoices for roundabout reasons in the slightly silly
name of Sonic Hedgehog. (audience laughs)
Don't ask. But this doesn't by any means imply that this Sonic Hedgehog protein, or the gene that encodes it, is a gene for villous growth. In fact, Sonic Hedgehog
is a general purpose, embarrassingly to some extent, it's a general purpose ingredient that keeps cropping up again
and again in development. And in this instance, what
happens is that in effect it has the effect of
switching the cell types in the epithelial layer so that some can keep on growing while others stop. And so this is basically what happens. You have this layer of tissue, and if some of the cells start emitting Sonic Hedgehog protein, if by chance a little bulge develops in that tissue as it grows, then this concentrates
the Sonic Hedgehog protein to a point where it
can trigger this switch and stop some of the cells growing so that they just continue
growing at the base of this unit and the rest of it just
then gets sort of pushed up further and further into this sort of finger-like protrusion. And it's a self-amplifying process. The more it gets kind of constrained, the better a trap it is for
the Sonic Hedgehog protein. So the real cause of these villi is therefore really a mechanical one. It involves changes to the
rigidity of the epithelial layer. Here's another example. We generally have five
fingers on each hand, right? But there's no gene that specifies that. Certainly, no gene that
specifies this number five. The way our fingers
are now thought to grow is that in the paddle-like bud of the developing limb in the embryo, there's a small set of, again, general purpose developmental proteins that interact with one another
in a particular complex way, a way that was first talked about by the British mathematician
Alan Turing in 1952. Turing showed how it's possible for a soup of reacting chemicals to spontaneously segregate into stripes of different composition, different concentration
of the ingredients. And that's what seems to happen in the development of the fingers. That stripes, radiating stripes develop in this bud of the growing limb, and these concentrations,
these stripes in turn trigger the growth of bone
that becomes the finger bones. And the reason that there are five of them is that the stripes have
a just an intrinsic width that depends on the
properties of the proteins. And they happen to grow at just the stage where five of these stripes will fit within the embryonic limb bud. Now, the whole growth process
is more complex than this. It always is in biology, but it seems that only a little tweak in the growth conditions
or in the timing of it might be enough to generate
more or fewer stripes. And that's possibly what we see in the way for ray-finned fish, they have more of these stripes that develop in their fins. So you can see here that
genes are providing resources for the body plan, but that plan doesn't
exist in any meaningful way within the genome itself. The key genes and proteins
involved in both these processes and in many others are, they're just general purpose
developmental proteins. As I say, they're not
proteins for developing any particular body type. The story isn't really about them as such. It's about the cells and tissues of the developing organism being triggered to make those proteins in just the right time
and place and sequence. And this shift in the location
of true causes in biology, it's not just a metaphorical
way that I'm using to talk about what's going on, it's something we can measure. When the primary causes of the behaviour of some complex system arise, not at the lowest levels
as they do for clockwork with all the cogs having
to fit in the right way, but that if they appear at
higher levels of organisation, that's something that scientists
call causal emergence. And we can see it all the
time in our social structures. For example, in the way a
company can still operate if some of the workers are off sick because there are usually
ways that the others can kind of compensate for their absence. And we can see it in the
flocking of birds, for example. So that if there's a kind of acentric or a tired bird somewhere in this flock that's doing something different, it doesn't cause chaos and
confusion among the whole. The larger scale organisation is robust against any little disturbances at the lower levels like that. And there are ways of measuring the amount of causal emergence in complex systems. And when a team of scientists
applied those methods to look at the mechanisms that are used by simple organisms like
bacteria, so-called prokaryote, and at more complex so-called
eukaryotic organisms like us, they could see a clear difference. We eukaryotes have more causal emergence, and I call this causal spreading. And it's spreading
rather than just a shift in where the causation is happening because the cause really is
spread across a range of levels. So there are still some traits, like diseases like cystic fibrosis, that can really be
pinned to a single gene. There are some traits that
are, in a meaningful way, caused by that gene. But most of our traits are
caused at higher levels, above the genes, even though the genes
can still influence them. Perhaps the ultimate expression
of this causal spreading is the brain. In "The Selfish Gene",
Richard Dawkins sounds almost affronted that our behaviour sometimes seems to go against
what a selfish gene picture should make us expect. But I think this is precisely
the point of a brain. You see, the challenges faced by bacteria and the decisions they have to make really aren't that diverse. You know, where's the food? Where's the moisture? How do I have to move to get there? They tend to live in single environments and they tend to die if
they go outside them. But we get everywhere, and every day we face
situations and challenges that we have never encountered before in quite that way in our lives, or that our ancestors have
ever encountered before. So, no genetic programme
is going to tell us what we should do in the
face of every eventuality. And the responsibility for that decision must be passed up to some
higher level to our brains, which don't have a kind of programme that is able to compute
exactly what we should do in any given circumstance, but is able crucially to improvise, in the face of the unexpected to improvise using fuzzy rules of thumb, not some precise, digital computation. And this way of behaving, so not through totally automated and predictable machine-like
stimulus response, but by genuine cognitive processing, this isn't just a good analogy for how life works at all levels, even down to the level of single cells. Some biologists argue that actually it literally is like that. That all living things should be genuinely
considered cognitive systems. As the biologist Mike Levin and the philosopher Dan
Dennett have put it, life is cognition all the way down. And this doesn't mean that the bacteria have some kind of mind worthy of the name, let alone any kind of awareness. Cognition doesn't have
to require consciousness to be genuine cognition. Well, however you feel about this way of thinking about life, I think it does capture
one important truth about the way life works and that is, that the best metaphors
for talking about it aren't ones that come from technologies, clockwork, or computers, but are metaphors that are
drawn from life itself. What really distinguishes living things from any machines that we've yet made is that they're not automata, but they have real agency. And what I mean by that is
that they're able to manipulate and alter themselves and their environment in order to try to attain
some self-determined goals. Now, when we recognise
that organisms have goals, we usually sort of say, "Well, the goals for all organisms are to survive and reproduce." Right, but while an awful lot of behaviour can of course be explained that way, I don't think it's enough. I'll hazard the guess that
the goal you set yourself in coming here tonight wasn't
about eating and reproducing. And if it was, I'm afraid you're probably gonna be disappointed, (audience laughs)
although who am I to say? But I think that we're
not the only animals in having agendas and purposes of our own that we decide that are
not obviously linked to evolutionary imperatives and aren't wholly predictable. Even single celled organisms and single cells of our body set their agendas to some
degree so that for example, what can seem like identical cells might behave in different ways to an identical stimulus because there is some internal
setting that they have that determines that. If you like, they've sort
of made up their own minds. They have their own goals. I think that displaying agency is actually a more fruitful
and more general way to think about what living organisms are than to try to come up
with some kind of checklist of, you know, what life is. Like say, reproduction, metabolism, homeostasis, and so on. That's why to have an overarching
view of how life works, I think that biology
needs an understanding and ideally, really, a theory of agency. What are the basic ingredients
that would require? We don't really know, although this new book by the
neuroscientist Kevin Mitchell makes a superb stab at
starting that conversation and along the way shows how it is that what we
call our own free will is really just an aspect
of the kind of agency that we, as complex, cognitive, and conscious beings, possess. But I do think that we
might be able to see some of the things that agents are probably going to require. For example, they need to make predictions about their environment
so that they're not constantly wasting energy coping with things that they might have anticipated and avoided. And in order to do that, an agent needs some kind of memory in which it can build
up and store information about its environment, which amounts in the end to a kind of representation
of a crude model of its environment. And we all have that. You know, if you leave here tonight and you were going back to Green Park tube and you start heading north
along Albemarle Street, then, you haven't really stored a good internal representation
of your environment because it's the other way, it's south. So you've wasted energy, which in that case is not
a matter of life and death, but sometimes making the
right prediction could be. Okay, this new view of how life works is I think it particularly matters when life isn't working so well and when we want to put it right. It matters for medicine. This issue of causal
spreading matters for medicine because if we want to affect
some change to a system, we do best to intervene at the place where that outcome is caused. Is the cause of such and such a disease some gene that we should be targeting? You see, what we tend to hear about in discussions of gene-based medicine are the exceptional cases, where cause really is situated
in a given identified gene. For example, in the recent announcements of the use of gene editing, gene therapy to treat sickle cell disease, which is primarily caused by
mutations to a single gene so that we can, in
principle, use genome editing to go into that gene and to put it right. But I think it's a fairly well kept secret that most of the regions of the genome that are found to be associated
with most common diseases aren't even within genes at all. They're found within
the non-coding regions that are presumably involved
somehow in gene regulation. And it's often very hard to
make effective interventions at that level anyway for one reason, because the genetic
effects tend to involve lots of very tiny effects spread across lots of different
regions of the genome. But I think ultimately,
the reason why it's hard is because the real
causes of these conditions operate at some higher level
of organisation than genes. For example, in the functioning
of the immune system. That's a common one. So we might still see
associations of the condition with particular bits of the genome, but in a sense all that
we are seeing there are weak echoes of the genuine causation sort of coming down from higher levels. That may be why genetic-based approaches to cancer treatments in particular have been so disappointing. You see, because cancers, we know that cancers
can arise from mutations that can happen to our genes, perhaps just because we get old or because we've been exposed to something in the environment that causes them. And so it used to be thought that we could find cures by looking at the genetic roots of cancer. But it seems increasingly
that the most effective levels of intervention are higher ones, and again, in particular, intervening in the immune system to help
the body itself fight cancer. As the cancer biologist
Michael Yaffe said in 2013, we spent fruitless years
looking for cancelling genes, not because we ever really had any reason to think they were the key to developing new treatments, but because we have the
techniques for looking for them. As Yaffe said, "Like data
junkies we continue to look to genome sequencing when the really clinically useful information
for cancer therapies may lie someplace else." This broader view of what governs
the behaviour of our cells and tissues and bodies,
I think also matters for understanding the kinds
of things that they can make. For example, we've discovered
in the past two decades that our bodies, that our cells can be tweaked to switch
between different states so that cells in our bodies
that have already developed into a mature tissue type,
a particular tissue type can be turned back and switched to a stem cell-like state from which they can develop into any tissue type. And I've experienced this myself directly. I've had cells of my skin
taken from my shoulder reprogrammed into that
stem cell-like state and then developed into neurons that grew into structures a bit like this called brain organoids, which look a little bit, not just sort of in
colour and size and so on, but actually in anatomy
look a little bit like developing embryonic brains. And this, looks like a normal embryo. Actually, this is like
a normal mouse embryo. They're made from mouse cells. But this is something a structure that has assembled
spontaneously from stem cells. No egg or sperm was involved in creating this so-called embryo model. So what our cells can make is not prescribed or pre-ordained. I think it's better to think
of them as being imbued with the potential to assemble into forms. And if we can understand what those higher level
principles of assembly are, then who knows what new forms
we might be able to make. But there's surely also good reason to understand these questions, to understand how life works simply to make us appreciate all the more how astonishing life is. The idea that what makes
living matter alive and different from a rock is simply that it's being
programmed to be alive is not just incomplete, but it's a bit boring. One of the sobering things about a gene-centered view of
evolution and of life is that it turns out to make a puzzle of why organisms exist at all. Richard Dawkins has called this
the paradox of the organism. Now, I don't know about you, but if my theory of to
explain the profusion of wondrous life forms ended up implying that those life forms
shouldn't really even exist, then I wouldn't so much call it a paradox as go back and think,
"Where have I gone wrong?" There are ways to think about this paradox within the selfish gene view of life, and it's interesting to do that, but I think it's not too
hard to see ultimately why this paradox arises. If you have taken all
of the genuine agency that exists in real organisms and squeezed it into genes to make them look like little
organisms in their own right, all existing and competing with each other in some in undifferentiated pool with an agency they don't possess, then it's not really any surprise that you no longer seem to
need real agents at all. The idea that an account
of life is to be found by this gradual increasing
of the magnification until we arrive at molecules is a fantasy, and that's something that the Nobel laureate biochemist
Albert Szent-Gyorgyi understood very well when he said this. He said, "My own scientific
career was a descent from higher to lower dimension, led by a desire to understand life. I went from animals to cells,
from cells to bacteria, from bacteria to molecules. On my way, life ran out
between my fingers." What it this comes down to then is looking for an
explanation of how life works that really does justice to the truly amazing
nature of life itself. Saying that life is just
a machine made by genes or a computer running a programme is not only wrong from the perspective of what modern biology is now telling us. I think it banishes life
altogether from biology. It sterilises it. The eminent physicist Michael
Berry told me recently that he was once asked the question, "What is the biggest
unsolved problem in physics?" And he figured that the question, it was probably expecting him to come out with some standard answer like dark matter or quantum gravity. But he found himself saying that, "If, as we think, all matter is described by quantum mechanics, then where does the aliveness
of living matter come from?" Now he didn't mean, thank God, that it must have some
quantum explanation, but he meant that living
matter is profoundly different from other kinds of matter and we don't really know why. Parts of the universe, all of these parts have become aware of themselves and their place within it. Biologists should never forget that what they're really trying to do is to understand and to explain that. No wonder, it's really hard. But I believe that we should refuse to accept too cheap an
answer to this question, the most profound question in science. Thank you. (audience applauding)
