Kat - Moving from plants to animals, Professor James Sharpe, at the Centre for Genomic Regulation in Barcelona, Spain is using mathematical models to understand how vertebrates like mammals and birds build a skeleton. I asked him to explain how he and his team start to model something as complex as this.
James - So, it probably is a complex thing. In our case, we're trying to address the very first basic steps that have to be taken so that it's not too complicated. It's a question of knowing how many bones to have in each position in your structure. So for example for the arm, we have one bone in the upper arm, we have two bones, the ulna and radius, in the lower arm, and then we have five fingers. Our research is to try and understand how that basic arrangement is laid out.
Kat - Now, there's an existing model of how this works and that you have a gradient of a chemical that goes across the developing limb. It goes from one side to the other and that tells you where the fingers go. But you're coming at this from a different angle. What's that?
James - So, we're actually trying to resolve the fact that there have been two competing ideas to explain this. The alternative idea to the one you mentioned was an idea that was first proposed by Alan Turing in 1952, although he didn't propose it specifically for limb development. It was a general observation about how cells might be able to organise themselves. But his theory which was a self-organising theory rather than a positional information theory has not been considered so relevant for this question over most of the decades of research in this field. We and others are starting to realise and do research to show that in fact, this self-organising idea of his is probably the main mechanism underlying the initial distribution of fingers in your hand for example.
Kat - So, in its purest form, how does Turing's idea work? What sort of patterns does it generate?
James - So, it always generates patterns that are periodic, which is to say they're alternating states...
Kat - Stripy...
James - ...stripy or spotty. In two dimensions, that's stripy or spotty. It can only do that because it does not have information about what's going on over the whole field of cells. It works purely by a local communication between neighbouring cells. For this reason, it can only produce stripy or spotty patterns. But this is exactly the kind of pattern that our fingers are - it's essentially a collection of five stripes next to each other.
Kat - So, if you just leave cells to do their own thing, more or less, within certain parameters, they will just organise themselves into stripes and you had a lovely quote in your talk about how Turing said that explains the stripes of a zebra, but not the horse bit. How is your work trying to understand the horse bit skeleton underneath?
James - So, in this case, as I say, the horse bit is the skeleton, but it is essentially just another example of a set of stripes. So, it is exact theory, unchanged, can be equally applied to do something like the fingers as to the stripes on the outside of a zebra. So, we're simply doing computer simulations, modelling and experiments in the lab to actually test whether his hypothesis is a reasonable explanation of that process.
Kat - But the sort of patterns that Turing's mathematics generates, they're not the straight, regulated digits that we see on a hand - they're kind of a bit blobby and a bit funny. What else is going on in the development of an actual hand to give that more regular pattern?
James - So this is where we realised that to actually produce something useful like a hand, rather than just a random Turing pattern, we actually needed collaboration between a Turing-type self-organising system, and also we need some kind of positional constraint that control this system, which could loosely be considered as a kind of positional information. So our general conclusion about how these systems work is that actually, rather than being a conflict between two alternative possible theories, both of these theories are equally necessary to explain the final result.
Kat - So we already know that there are lots and lots of different genes that are turned on in the hands, in the limbs, in different places and at different times. So it's the interaction of those with the underlying pattern that's shaping our bodies, literally?
James - This is exactly what we think. We're still probably at early days and we have a lot further to go, but it is exactly those gene expression patterns, signalling pathways, gradients, which regionalise the tissue, which have different values in different places, which control the self-organising process.
Kat - In your talk you talked about the long bone of the arm, the two bones of the lower arm and then the five fingers. What about the wrist? There's loads of bones in there!
James - That's a very good question actually. It's been speculated that it might relate to something that was discussed in connection with Turing more explicitly, which is the difference between zebra stripes and leopard spots. So you can change a parameter of a Turing system to switch it from making stripes to making spots. In fact, the wrist bones are more like spots, and the other bones are like stripes. So it could well be that the parameter is tuned just in the wrist region to convert from stripes to spots.
Kat - So with the models you're generating, understanding the stripes, the spots, the parameters that they're tuned within - where ultimately do you want to take this work? To figure out how we can build a whole skeleton?
James - Well, in fact our interest is not to explain the skeleton of the rest of the embryo, but instead to focus on the limb and to bring in the other factors that are relevant for the limb. Our goal is to make a complete computer simulation or model of all the relevant important aspects of limb development. So we're dealing with the skeleton at the moment. We're also dealing with mechanical morphogenesis of the bud, we're also trying to understand the regionalisation - why the hand is different from the arm? And our real goal is to bring all of these together into a single computer simulation where we can then understand, potentially, the higher level more complicated aspect of how these different subsets, subsystems interact with each other to create the full organ. The real ultimate goal is to explain organogenesis as a whole.
Kat - Now we more commonly hear Turing's name mentioned with regard to his efforts in code-breaking and things like that. Many people don't know that he had these ideas about biological systems. Do you think more people should be aware of it?
James - Personally I obviously think more people should be aware of it. In fact, I went to the centenary meeting of Turing last year in Cambridge and even there I was shocked to discover that the vast majority of researchers that had come to the meeting were completely unaware of his contribution to this area of developmental biology, and in fact it was an area of mathematics that he contributed to. It spawned an entire new field of mathematics, and yet still - as you say - many people are not aware of his contribution.
Kat - And now, sixty years later, you're trying to bring it back?
James - Absolutely!
Kat - That was Professor James Sharpe from the Centre for Genomic Regulation in Barcelona.