Digital Digital Burgess Conference Follow-up:
Richard Gordon
Conference was held August 29-September 1 1997, Banff Alberta, Canada

More on differentiation trees and how they relate to embryonic development, genetics, and evolution

The differentiation tree seems abstract, but let me try to explain it. In an organism, each cell comes from the division of a previously existing cell. (Muscle is sort of an exception, since the fibers are formed by fusion of many cells together into what is called a syncytium. But we could think about their nuclei instead, which retain their identities.)

This hierarchical arrangement means that all the cells in the body are related by a tree. At the base of the tree is the one cell embryo it started from. This tree is called the cell lineage tree.

Now, in our bodies, many cells are of the same kind. Suppose we gathered and bound together all the branches of the lineage tree that correspond to the same cell types. It's still a tree, but with many bundles, one for each differentiated cell type. If we replace each bundle by a thick line, then each line represents one kind of cell. This bundled tree is called the differentiation tree.

Consider one branch point in the differentiation tree, say A - B+C. This means that we have a group of A cells at a given stage of embryonic development, and that group splits into two groups, each cell of which becomes a new cell type: B or C. (If no further cell types develop from, say, the B cells, then we call them "terminally" differentiated cells. Note that A cells don't exist in the adult, since they've all turned into B or C cells, and whatever those may have turned into. Of course, cell division makes more cells of each type as the process goes along.)

What we've observed is that a visible, physical wave goes through a portion of the A cells, turning them to B cells, and then another wave travels through the remaining A cells, turning them into C cells. These waves are contraction waves and expansion waves. So the nodes of the differentiation tree (where the lines join) each represent launching of a pair of differentiation waves.

If these observations hold up as a general theory, then the major problems we have to solve are:

  1. What launches a wave from a particular place in an embryo at a particular time?
  2. What shapes the wave's trajectory as it moves across a tissue?
  3. How does a wave trigger specific gene expression in the cells it travels through?

I'm working on these problems, but am rather short handed.

Now, as to the brain, my suggestion is that it has its own major branch of the differentiation tree, and that branch is huge in primates, especially us. This suggests that the brain is constructed as a hierarchically divided set of regions of the original brain tissue (= neural plate or neuroepithelium). If right, this idea might help us understand and perhaps design "brains". It may be that conditions like schizophrenia represent incomplete formation of one or more branches of the brain portion of the differentiation tree.

The most spectacular wave we've found so far is the one that leaves the neural plate in its wake. We call it the ectoderm contraction wave, since it passes through one of the two hemispheres of the tissue called ectoderm, turning it into neural plate. (The wide, head end of the neural plate later turns into the brain, while the narrow, tail end becomes the spinal cord.) The ectoderm contraction wave starts at a point, breaks into an arc when it hits another set of waves at the dorsal lip of the blastopore. The ends of the arc travel faster than the middle, so that the wave (a single, travelling furrow) closes up as an ellipse and then vanishes as the ellipse shrinks at what will be the head end of the embryo. The speed is about 3 microns/min. ( 1 micron = 1/1000 mm = 1 millionth of a meter) The wave takes about 12 hours from start to finish in a 2 mm diameter axolotl embryo.

To relate this to the Burgess shale, my point was that each kind of organism had its own, unique differentiation tree (which I identify as the Bauplan = German word for "building plan"). Furthermore, a major theme of all evolution above the one cell level is that of the evolution of differentiation trees. This probably happens by copy/cut/paste of chunks of DNA bearing the genes for differentiation. The genes expressed at each branch of the differentiation tree (i.e., at a particular step of differentiation) have to be located somewhere on the DNA. Thus there is some kind of map between the DNA and the differentiation tree. I suggested that groups that succeed in radiating (forming many species occupying many ecological niches) are precisely those that have a simple map between the DNA and the differentiation tree. The present day mammals represent a relatively recent radiation.

The simplest way for the differentiation tree to grow over evolutionary time is by copying short, terminal branches of the differentiation tree via DNA copy and past operations (multiple "gene duplication", as there are 50 to 300 genes per step of differentiation). If something about this operation favors growth over pruning of the differentiation tree, then evolution is automatically progressive. Exactly what this "ratcheting mechanism" is would be a 4th major problem to be solved. At the conference I came up with a simple possible mechanism for this: gene duplication is slightly harmful, so duplicated branches will generally evolve to become different from one another, increasing fitness by: 1) eliminating the double dose of identical genes; 2) forming new, different tissues that permit greater complexity in the organism. Tom Ray liked that idea, so maybe some of these concepts will find their way into his Tierra.

Thus differentiation trees tie together embryonic development, genetics, and evolution. Hence the title of my forthcoming book:

Gordon, R. (1997). The Hierarchical Genome and Ultraslow, Differentiation Waves: Novel Unification of Development, Genetics, and Evolution (Singapore: World Scientific), in prep.

by Richard Gordon

Department of Radiology
University of Manitoba, Health Sciences Centre
Rm. A246, 820 Sherbrook Street, Winnipeg, MB R3A 1R9 Canada

Conference Home