Richard Dawkins and Denis Noble discuss Darwinian evolution, selfish genes, and whether animals are active agents or merely biological machines.
Can organisms change their DNA?
Dawkins' Selfish Gene has been hugely influential, both within evolutionary biology and in the wider public sphere. It's a beautifully simple story: genes and not organisms drive evolutionary change. But critics argue the story is simplistic. The effect of a gene is not always the same and as is dependent on its host and the cell environment. DNA does not come neatly divided into individual genes. And in 2010 the renowned biologist EO Wilson and others revived the case for group selection. Some are now arguing that the Selfish Gene paradigm is holding back medical research.
Is it time to move on and acknowledge that Dawkins' theory is not the whole story? Might his theory be making a fundamental mistake in reducing humans to machines? Or does the Selfish Gene remain a remarkably powerful and accurate account of who we are? Richard Dawkins is a British evolutionary biologist and author. He is an emeritus fellow of New College, Oxford and was Professor for Public Understanding of Science at the University of Oxford from 1995 to 2008. An atheist, he is well known for his criticism of creationism and intelligent design.
One of the founders of the field of systems biology, Denis Noble is an Emeritus professor of cardiovascular physiology at the University of Oxford. He was the first to develop computer models of the heart in 1960. His books include the first popular science book on the subject, The Music of Life (OUP, 2006), which has been translated into 12 languages. Central to Noble's work is that living organisms are "more like a piece of music than a genetically determined machine", a challenging metaphor in the field for the field of biology. Amongst many awards and honours, he became a founding Fellow of the Academy of Medical Sciences and was awarded a CBE in 1998. In 2022, he was elected a Fellow of The Linnean Society.
Hosted by Güneş Taylor.
00:00 Organisms choose which of their genes will survive 00:37 Can the body change the genome? 03:58 The remarkable process of DNA selection 06:58 Problems that arise with Lamarckian evolution 08:53 Germ lines, gene mutation, and the pandemic 10:21 How many generations of mutation before a gene is assimilated into the genome? 15:05 What modern microscopes reveal about natural selection 17:15 Understanding Lamarck and Darwin 18:45 What gets passed on through generations? 20:43 Starvation effects, individual mutations, and changes in the gene pool
I think what organisms are doing is partly through their social choices,
effectively choosing which genes they will allow to survive.
That's what Waddington is doing.
Social choice?
Social selection, yes.
How?
Well, who you mate with, for example.
Why do we marry anybody?
Isn't that why we do it?
This is perfectly Darwinian, what you're talking about.
Yes, absolutely, I agree.
Now, I think what we need to do here is to get another element into this,
because I think what you're really worried about is
how can it be that the body can actually change the genome?
And that's the big question.
Now, the reason we know that it can is that we know it controls it.
That's the first step.
So let's see, first of all, how that can be done.
I have two very important colleagues who have done the work I'm going to describe,
so I'm going to credit them.
Dick Chen worked with me as a graduate student way back in the 1960s
and is now working at the University of New York
and has done part of the experiments I'm going to describe,
and Anant Parekh, who is a physiologist in the same department as me in Oxford.
And what they've done is absolutely beautiful.
They've asked the question, you see, it's the relevant question that I think Richard is asking.
How can it be that the surface of the body or of a cell,
it might be that it's a unicellular organism,
then it would be the surface of the organism.
How can it know, how can its nucleus know that there is a need to change?
And we now know how that can be done.
What they've shown is best described by imagining, first of all,
that a single nucleotide is about the size of my fist
and it's situated in the nucleus.
So let's put that in the centre of the cell.
If we did that, on that scale,
the surface membrane of that cell would be way up in Scotland.
How on earth can it be that a signal through a receptor on the surface
can influence the nucleus?
And we now know how that can be done.
What they both found, doing different experiments in different cells,
was that calcium coming through protein channels in that surface membrane,
using the same metaphor, way, way up there in Scotland,
creates a calcium concentration
in a small subspace underneath the membrane
and that high calcium triggers a chemical reaction
that produces a messenger.
And that messenger gets attached to some extremely important proteins in the cell.
Those proteins are called tubulins
and the name suggests what they do, they form tubes.
Literally, there are tube trains in cells.
And I'm not joking,
because what happens is those tubulins run all the way through
from one edge of the cell to another.
They have little motors on them, little molecular motors,
and they can attach a messenger molecule to the motor.
And what then happens is phenomenal.
They literally walk along the tubulin.
It takes just a few seconds to go from that surface,
imagine on this scale, way up there in Scotland,
to the nucleus.
What does it do?
In those experiments, it changes the gene expression levels
in the relevant genes that matter for that particular function.
Now, the only thing that's missing here,
and I'm sure Richard will pick this up very quickly,
so I'll say it myself,
is that those are very recent experiments,
done 2016 and more recently, 2018, I think it was.
Anyway, the important point is that we don't yet know
how that induces genome change,
and I really mean actual change in DNA.
And yet, we know also that those processes must be able to do that,
because we can show that.
Let's take a tumor developing your body,
and it's a bad situation.
You're beginning to get metastasis.
So the doctors get out the radiotherapy and the chemotherapy,
they attack it and try to destroy it.
What happens?
The tumor cells themselves tell the genome to increase the mutation rate.
How can they do that?
Precisely by the kinds of mechanisms I've just described.
Because the mutation rate is under the control
of what is happening in the body as a whole.
What then happens is phenomenal.
It happens in your immune system all the time.
It happens in bacteria all the time,
because they change their genomes in response to antibiotics.
And what they do is very simple.
You remember that difference between 1 in 10 to the 4
and 1 in 10 to the 10?
That depends, as I said, on the cell.
Having these repair mechanisms, the proofreading mechanisms,
but you see they can be down-regulated.
That process can be down-regulated.
And what that does is to produce literally millions of new DNA sequences.
That can then be selected.
Now the selection, and I agree there is a kind of natural selection here
within the organism.
Now the question is very simple.
Do those new sequences get to the germline?
You bet they do.
And that, I'm afraid, is where I think the big hole in the theory lies.
Because once you can do that,
you can get what, for example,
Zhang and his colleagues have shown in a paper published in 2018.
I can send all these references to anybody who sends me an email.
So if you're worried about whether I'm telling the truth,
just send an email and I will send you the reference.
What they showed was that a small non-coding RNA,
that's a little bit of technology,
but a new sequence generated by the organism
can pass to the germline cells,
which become eventually, of course, the eggs and the sperm.
And what that will do is then tell the next generation
to inherit the metabolic characteristics that were conveyed by that.
I'm sorry to say this,
because I know this is a dirty word amongst most evolutionary biologists,
but Lamarck is back.
Very simple.
All right.
By the way, the walking mechanism is simply beautiful.
It is.
You can see films of it.
It's absolutely uncanny.
At one point, Dennis,
I thought you were confusing gene expression,
which, of course, is obvious.
I mean, it happens all the time.
No, it didn't confuse them.
With germline changes.
That's why I went on to explain how those changes
can then be communicated to the germline.
That's an extremely important distinction.
Indeed, yeah.
There's no dispute whatever about certain genes being turned on
in some cells and others in other cells.
That's what embryology is all about.
However, what Dennis went on to say
is that there's evidence that it actually gets into the germline
and Lamarck is back.
Well, if Lamarck is back for an indefinite number of generations,
I'm impressed.
If it's only for a couple of generations, I'm not.
But let's suppose that it is for a larger number of generations.
If that's true, then I would have to revise what I say
to include any change in the germline
that now becomes admitted into the charm circle of replicators.
And that's fine.
Well, it would, yes.
I doubt it.
I know you wouldn't.
But I don't want to be dogmatic about saying
that the DNA in the existing germline is all there ever was.
If on some other planet, and maybe on this planet it's true,
the germline can be altered, then that's fine.
The broad church of the selfish gene can embrace that.
As I say, I doubt it.
Yes, but look, Richard, I think one thing to perhaps make clear
to the audience is this is happening in everybody in this room
because we had the pandemic that arrived with coronavirus.
Now, of course, we've fortunately developed vaccines
against the virus, and that's been our great saving grace.
But what would have happened anyway,
with a lot of people dying, of course,
would have been that our immune systems would have done
exactly what we're describing.
That is, they would have used that mechanism for hypermutating,
that is, mutating extremely quickly to produce millions
of new DNA sequences.
And then that is used to be what then gives you the immunity,
the acquired immunity, obviously.
Now, what Richard is questioning is, okay,
maybe that can occasionally be passed to the germline.
We don't know that yet, whether an immune response can be passed
to the germline, and I would readily say we don't know that yet.
But what is important is Richard's point about how temporary it is.
Now, it's very important indeed, and I agree with Richard
about the importance of temporariness or permanence,
because it seems to me that what these mechanisms give
is the option for the evolutionary process
to, as it were, try it out.
If there's an environmental change that makes it very difficult
to survive, and all organisms are under stress,
and they alter their genomes and pass some of that,
even temporarily, on to the next generations,
what the next generations can do is to find out whether
they do experience that change in environment or not.
If they don't, then it's great that it's temporary.
You don't have to alter the main genome.
If it is more or less permanent and goes on for many generations,
then how can it get assimilated in the genome?
Conrad Waddington showed how to do that way back in the 1950s.
Incidentally, his book, The Strategy of the Genes,
has been rightly republished in 2014, so you can buy it again.
It was published in 1957.
He did beautiful experiments on fruit flies.
He induced changes with very tiny, gentle persuasion, as it were,
from either heat or ether or some other experimental techniques
in which he could, as it were, persuade a few of the fries
to show a new characteristic.
And he actually determined how many generations
would you have to continue to do that
in order for it to become assimilated into the genome.
It's about 14. It's not very long.
Now, what he was showing is what he called genetic assimilation.
I think it was a great mistake that Waddington was ignored
by the evolutionary biologists, and that's a shame.
The Waddington effect was actually selection.
I mean, it was not Lamarckian.
By him.
Well, the flies that didn't respond correctly to the heat shock died.
Yes, that's right.
So it was selection. It was Darwinian.
I'm agreeing with you.
It only looked like Lamarckian.
This may be the only point in the evening where we totally agree.
That was selection. Yes, I absolutely agree.
I'm agreeing with you.
What Waddington was doing was a simulation
of a Lamarckian experiment for quite a different reason.
And I think it comes back to your opening question to me.
Do you still hold to the idea that it's agency
that organisms have rather than the DNA?
No, I do.
Because, you see, I think what organisms are doing
is partly through their social choices,
effectively choosing which genes they will allow to survive.
That's what Waddington was doing.
Social choice.
Social selection, yes.
How?
Well, who you mate with, for example.
We're back to Darwin's idea of sexual selection.
Well, we are.
The social selection ideas.
So why drag Lamarckian then?
I think that's Lamarckian
because it's part of the use within the social context.
You see, what Lamarck was insisting on
was the idea that use and disuse
was itself something that could be inherited.
And I think this is something, of course it starts culturally,
but it becomes something that can be inherited
through the fact that you are, as organisms,
choosing the characteristics
that you want to survive in the later generations.
Why do we marry anybody?
Isn't that why we do it?
But, Dennis, you're...
I mean, this is perfectly Darwinian,
what you're talking about.
Yes, absolutely, I agree.
And Darwin was a Lamarckian.
I'm not joking.
No, you're not.
No.
In 1868, he published his theory of gemmules,
which is precisely the thing we've now discovered
as the extracellular vesicles today.
So I absolutely agree with you, Richard.
Darwin was indeed a Lamarckian.
I'm a good Darwinian.
You're a 6th edition Darwinian.
Darwin, in the 6th edition of The Origin of Species,
did flirt with Lamarckism, that is true.
That's a historical fact,
but it's not a very important biological fact.
Oh, I think it's extremely important.
OK, well...
No, seriously, Richard, because he collaborated with...
This is not very well known.
He collaborated with physiologists
in the last 20 years of his life,
between 1872 and 1882.
He collaborated with my predecessor
as the chair of physiology,
Burden Sanderson,
and he collaborated with his student,
George Romanes,
in a very simple set of experiments,
because, you see, he took Lamarckianism so seriously
that he invented this theory of Gemmules,
and I'd better just very briefly explain what that is.
He realised, as Richard has beautifully explained,
that you've got to explain how it can be
that the body can, in its changes
due to use and disuse,
communicate any of that to the germline.
Otherwise, all of that information,
as Richard beautifully expressed it earlier on,
would be lost.
So how can that be communicated?
He couldn't see what could possibly do that,
so he invented an idea,
and he admitted it was an idea,
which was that tiny particles
put out by the cells themselves,
which he called Gemmules,
would be able, perhaps,
to pass through the bloodstream
down to the germline.
That was his way of explaining
there could be soma-to-germline expression.
But he readily admitted at the time
this was just a hypothesis,
because he couldn't see them.
Now, with 19th century microscopy,
indeed you could not.
With 20th century microscopy,
and 21st century microscopy
even better, we've been able
to do so, and the experiments
are simply beautiful. Just go online
and ask to look
at extracellular vesicles
made evident
by labelling
molecules fluorescently,
so they literally glow
green, yellow, red,
or whatever it might be. It enables you to know
this is this particular RNA,
this is this particular DNA, and so on.
And that escapes
the limits of light microscopy.
You can actually resolve
down to very tiny particles
indeed. They're called
extracellular vesicles.
Those have been shown experimentally
to be passed to the germline.
That's how the RNAs and DNAs,
the new RNAs and DNAs, get
to the germline.
So, I think that
if he was alive today,
I think Charles Darwin would be praising
and cheering
the discovery of extracellular vesicles.
They are his
gemmules, and they carry out the function
that Darwin proposed. Now,
why did he spend the last 20 years of his
life collaborating with George Romanes
is because he actually thought this must
be right. So, I don't
think it's trivial that Darwin was
a Lamarckian.
I do think this is actually quite misleading.
What Darwin's gemmules
were supposed to be about
was investigating
the current
state of the body and passing it on to
the next generation. So, the gemmules were going
all around the body, and they were
detecting
changes in the body.
The sort of classic Lamarckian
examples like the blacksmith's arms
getting muscular and the giraffe's neck
stretching and things like that.
Lamarck thought
that those were inherited. Darwin
in his later years thought
they were too,
and Darwin's
gemmules were going around the body in the
bloodstream and picking up
information about the
current state of the body, the modified
state of the body, the acquired state
of the body, and going to
the germline, going to the gonads,
and imprinting the information
into the germline.
Now, that is a very radical
idea. It's precisely what the extracellular
vesicles are doing. Well, yes,
but they're not.
It's nothing to do with blacksmith's arms.
They may be
doing something. If you're right about the immune
system, you seem to be suggesting
that what happens is that
when the immune
system
reacts to
an infection like COVID
and
we become
immune to it, that immunity gets passed
on. No, I deliberately
said we're not yet sure about that. I know you
did, and I'm glad you said that. But what we are
sure about is that other things are
passed on. Metabolic disorders
are passed on, and sexual
preference are passed on. Well, sexual
preference is in what
way passed on? It's passed on
in planarians, and that's
been demonstrated. Again, all of
these references, I'm very happy for people
to email me and ask for them,
but that's been shown very
recently by Toker and his collaborators
in work in Israel,
and I think that is actually a 2021 reference.
And how many generations?
Well, what they're showing,
okay, come back to the point
I made about temporary and permanent,
because, you see,
temporary is actually an advantage
if you don't yet know
from an evolutionary perspective whether
the change is valuable or not.
I think it's great, you see,
that epigenetic changes
and temporary alterations
of the germline
are not
necessarily passed on through
many, many thousands of generations,
because if the change in the environment
is really temporary, you
don't want a permanent response.
So I can see the evolutionary logic
of doing it in that
kind of way. You keep it soft
until it needs to become hard,
and then you let it become hard.
You let it then become assimilated
into the genome.
Well, that's fine. I mean, that's coming
back to the Waddington effect, in a way.
To some extent, Richard. Yes, I think
this is why I said that Waddington was
badly ignored.
Or sometimes called the Baldwin
effect. Sometimes called that.
Yes.
But I think what's
happened today is that we actually now know
the precise mechanisms by which it can
happen. We know the molecular biology
of it. We know the cellular biology of it.
So what I'm saying is
it's time for evolutionary biology to
catch up. I mean, if I may ask
in that case, how long would it need to be?
I mean, you've asked a few times. I'm really
taken by this sort of temporal thing.
How long would it need to be to have an effect?
Do you think?
In order to be evolutionarily
interesting, then it
needs to be something
that we see as a change
in the gene pool.
And
change in the gene pool would be,
I mean, I can't put an actual number of generations
on it, but
it's not a
proper Darwinian change
if it's just,
for example,
there's evidence that
starvation effects can...
I was going to ask about that next, yes.
These are,
as it were,
I mean, epigenetic
effects are
changes in the,
as the embryo develops,
changes in the expression
of genes in different parts
of the body. So in liver cells
certain genes are turned on.
In kidney cells other genes are turned on.
Muscle cells, other genes are turned on.
Those are epigenetic effects.
Now there is some evidence
that those epigenetic changes can be
inherited into the next generation
and possibly the grandchild generation.
That's not a
proper gene pool change.
Yeah, I think Richard is right on
that. But what we would need to do
is to look at the effects after
billions of years. And that's
exactly what the Human Genome Project
did in its Nature paper
of 2001. Remember I referred
to figure 42 of that paper.
You see there the evidence
that those genomes
were changed by
moving great chunks around
in the genome. Not
time I guess to go through the detail
of that. Yeah, unfortunately not.
That's fairly clear evidence that it must
have happened during evolutionary
timescale.
It's obviously also important to say that increasingly
with modern technologies people are starting
to look at sort of the genomes of
other humanoid species
and looking into the past
to sort of get more information
on perhaps what our more recent
ancestors look like. And it might be quite
interesting to sort of see
whether or not those pieces of data
can add to
this conversation in due course.
What bothers me is
Dennis is saying Lamarck is back
because in order for Lamarck
to be back it seems to me we would need to have
something more like the blacksmith's
arms effect. Where
an adaptation, and I mean there's plenty
of adaptations that happen in lifetime.
Your muscles develop
when you use them. It would
be wonderful, maybe on some other planet
it happens that when
your muscles develop,
you get a suntan,
all sorts of adaptive changes
like that get inherited.
And that's what
Lamarck was suggesting.
And I think to say that Lamarck is back
is going to give a misleading
impression because people will think you're saying
that something like the giraffe
effect, the blacksmith's arms effect.
To be very precise,
it is that the inheritance of use
and disuse is now evident.
That's the way I would see it.
Is it? I mean
you're not going to go out and say
that
adaptation as we see it in the
field, as animals develop
camouflage, as animals develop
stronger bones
as they use them, stronger
muscles as they use them, that would
be a proper Lamarckian effect.
That would be a real
adaptive change as a consequence
of use and disuse.