The enduring power of a well-formed analogy.
The human egg is an astonishing thing. Over the nine months that follow its fertilisation a single cell is transformed into an entire, functional, human being. The several billion cells that emerges mewling into the world are organised into a complete set of working organs: a nose, a liver, two kidneys, even a tiny appendix. The child has wired its own nervous system to begin the slow process of mastering its new body and making sense of its place in the world.
Throughout time scientists have pondered this marvel of self-construction. Aristotle was the first to write about the subject in his book On the Generation of Animals, back in the fourth century B.C.. The twentieth century scientist who did most to put developmental biology on a firm theoretical footing was Conrad Waddington. Recently described as “the last renaissance biologist“, Waddington was a unique character. With an academic background in philosophy, palaeontology, embryology and genetics, and a wide ranging engagement with the arts, he was, by anyone’s reckoning, something of a polymath.
One of Waddington’s most enduring gifts to biology is not a fact, hard-won through an elegant series of experiments. Instead it is an analogy, derived from an understanding the data available to him, and an intuitive notion of how embryonic development should proceed.
Waddington realised that the key to understanding the building of an embryo was to see it as a gradual process of specialisation. Through its life, each cell undergoes a series of choices. With time the choices become more specific. Most are irreversible.
To visualise this process Waddington conjured up a hypothetical landscape, his epigenetic landscape. The cells of the early embryo can be thought of as balls at a high-point in a craggy and eroded landscape. As the balls roll downhill, they will enter a system of branching valleys. There are many possible routes down the hill. Each fork in the valley system represents a decision point for the cell: Inside (endoderm) or outside (ectoderm)? Foregut or hindgut? Pancreas or liver? Insulin producing cell or duct cell? and so on.
The picture is clear and visually appealing. But at first sight it is not completely clear what the ball represents, and what is the ground over which it travels? In an illuminating series of letters exchanged over the summer of 1974 Francis Crick, Nobelist and co-discoverer of DNA’s structure, demands some clarification from Waddington.
Crick kicks off the dialogue on the sixth of June 1974, nearly 35 years since Waddington first described the epigenetic landscape:
“I have been wondering recently about your “epigenetic landscape” and have to admit to having some difficulty in deciding what you meant by it? … of course one grasps in a rather vague way what you mean to convey, but I have found it difficult to make the analogy precise.”, Crick writes.
Waddington replies immediately and at some length, claiming
“It is a simple and perfectly clear idea.”
He explains that the ball represents a mathematical “coefficient”: a number that sums up all different components within and around the cell that influence its developmental decisions. He states that
“the components will be chemical molecules, the most important of which are directly (and the others indirectly) specified by the genes.”
In other words, the action of the genes control the contours of the epigenetic landscape, and hence the ultimate fate of each cell.
In his 1957 book The Strategy of the Genes Waddington depicts the the genes as massive subterranean ropes that marshall and pull the surface of the landscape. “The ropes are coupled into a network, so that a change of tension of one rope might affect wide areas of the surface”, he writes. In the 1940s and 50s when Waddington was first formulating his landscape analogy, his was a startlingly original view of gene function.
In his letter to Crick Waddington is at pains to stress that
“the very numerous genes will interact in such a way that will produce only a smallish number of distinguishable cell types”.
This is a key feature of the landscape model: only a finite number of cell types can ever arise. In the parlance of modern-day complexity theory, cell states represent stable attractors in an unfathomably complex multi-dimensional system.
Later in June Crick replies, demanding further explanation:
“Can one imagine an organism evolving in a way which could NOT be symbolised by the landscape? If you can’t, I don’t see that the analogy has any force.”
A note of tetchiness creeps into Waddington’s reply:
“You seem to be making such heavy weather of grasping the point”.
He goes onto describe how his landscape allows the influences on the cells of a developing embryo to be dynamic and change continually with time. He imagines a more static system where development would unfold according to simpler Newtonian rules:
“the landscape is then as flat as a billiard table, and of no interest.”
Crick remains unconvinced and challenging.
“This may have been useful idea in the Thirties, but I think it has long outlived its usefulness”, he writes. “In any case, why a landscape? Why not a railway marshalling yard, with tracks repeatedly branching from a single track?”
Waddington becomes increasingly exasperated.
“In your current mood you do not intend to understand the epigenetic landscape, and one might say there is no point in discussing it further. But I feel that, for purely formal reasons, I should not leave you talking such nonsense without putting some reply on record”.
He explains that the three dimensional rendering of the landscape is crucial. It fits with what developmental biologists were learning from their experiments. Some cellular decision points are abrupt, whilst others occur gradually. In the same way the landscape can contain both sharp ravines and gradual inclines. The difficulty of the terrain separating two valleys also indicates how hard it is for a cell to switch track; how stable its attractor state is.
The final letter in the exchange comes from Crick, who strikes a conciliatory note at last:
“Peace! Peace! I really am trying to get the most out of the epigenetic landscape, even if at times my manner gets a bit too brisk.”
The pair planned to meet to thrash out their differences in person. Whether they managed that hasn’t been recorded and Waddington died just a year later. Even if they did meet, I doubt they would have settled the matter. The letters illustrate the collision of two great minds. But also two very different minds.
On the one hand Waddington, the polymath, the generalist, paints with a broad brush. He wants to capture the whole sweep of theory, and create a vivid image that conveys the essence of the biological process. One the other hand is Crick, the intellectual surgeon. Also searching for the fundamental patterns, Crick wields a scalpel to pare down and trim excess fat from the theory. He demands absolute specificity and direct applicability to real living systems.
Crick’s most famous model, the double helix of DNA, is a different beast from the epigenetic landscape model. The helix is a minimal description of a collection of focused experimental evidence. The epigenetic landscape attempts to unite a much wider body of data and theory, and provides a different type of summary. It is important to remember that both models are still abstractions. In reality DNA isn’t always in a neat, linear double helix. In reality the DNA is deformed, looped, manipulated and chemically modified. The helix model ignores the grey areas. With the cost of being much less precise, the epigenetic landscape model welcomes ambiguities and celebrates the dynamics of the systems it describes.
The ultimate test of an idea is whether it has stood the test of time. The double helix has definitely passed that test. It set the direction of the whole enterprise of molecular biology. But as a metaphor and a tool for visualisation, the epigenetic landscape is also alive and kicking. It has influenced generations of biologists, locally and globally.
In Edinburgh, where he spent the last three decades of his career, Waddington’s vision struck a chord with many of his contemporaries. Through the 1960s and 70s Waddington was keen to usher in the new science of molecular biology. His ambition to understand the taxing problems of developing organisms required a detailed understanding of the action of the genes. The molecular biologists held the keys to that understanding. But at the time most molecular biology, including work done in Crick’s world-leading Laboratory of Molecular Biology in Cambridge, focused on bacteria and viruses. Partly under Waddington’s urging, many Edinburgh scientists picked bigger fish to fry. Edinburgh became an important centre for molecular research into larger and more complex lifeforms. The epigenetic landscape model didn’t solve their day-to-day problems in the lab, but perhaps it served as banner under which they rallied?
More globally Waddington is one of the intellectual fathers of the systems biology. With his emphasis on theory and his prescient grasp of genes working in networks, Waddington’s work feeds directly into this new discipline. Today’s systems biologists argue that a ruthlessly reductionist approach to biology is doomed to fail. Instead of looking at each cellular component in isolation, we need to zoom out and contemplate the properties of the system as a whole. How is information stored, managed and transmitted to control the identity and function of our cells?
Seventy five years on system’s biologists are still using and testing Waddington’s epigenetic landscape model. According to their latest mathematical models, many cell state choices don’t in fact look like a branches in a valley system. A world where valleys suddenly disappear from the map might be a more accurate rendering. Whilst it may not be correct in every detail, it’s value as a visual and immediate analogy is hard to dispute, if difficult to quantify. One thing the epigenetic landscape does beautifully is to provide common ground between the theoreticians and the experimentalists. This is crucial since the former are totally reliant on the data generated by the latter.
By bringing these two groups together the epigenetic landscape is one tool that is helping biologists to make sense of the miraculous unfolding of cellular potential that leads to the formation of every new human. As we understand that process better, we can start to do more to help when development goes wrong. We can also hope to re-purpose the tricks of embryonic development to repair damage and treat disease.