In developmental biology, an individual cell in a large population needs timing and positional information to participate correctly in the formation of patterns (morphogenesis) and in the determination of organ size. We explore the possibility that much of the needed timing and positional information is contained in a cell’s time‐evolving geometrical context, which it senses as strain cues. Following a top‐down strategy, response functions that map strain cues onto actions such as cell motion, cell shape change, positive velocity feedback, or secretion are built into simulations that address the inverse problem: what strain‐cued single‐cell response functions (if any!) can enable a population to match patterns, size, and other characteristics of some epoch in development? Successful response functions (there are a few out of many tried!) can all be viewed as either positive or negative feedback mechanisms, acting to augment or moderate a sensed strain variation. In simulations of evolving cell populations, sensed strain variations may reflect the activity of, e.g., 10 cells immediately surrounding the sensing cell that might offer, e.g., directional guidance to the cell’s motion, or 100 – 1000 cells including the sensing cell that are collectively determining morphogenesis in a small region of an organ (a “morphogenetic field”), or 100,000 – 1,000,000 cells including the sensing cell that are collectively determining the size of the entire organ via strain waves echoing off the boundaries of the population.

The theory contains neither mass inertia nor elastic strain energy and makes no reference to chemical factors. For pattern formation, the theory is an alternative to Turing’s reaction‐diffusion mechanism. The strain‐cued feedback theory is successful in as far as it can reproduce a number of patterns, including sliding layers of cells observed during amelogenesis (enamel formation), periodic or “segmented” patterns observed, e.g,, during amelogenesis or somitogenesis (the formation of vertebra pre‐cursors), and closed‐loop networks observed during innervation of the gut; and predict the size and shape of the enamel organ. Perhaps surprisingly, simulations make a number of quantitatively correct predictions (within reasonably expected uncertainty) whose number exceeds available adjustable parameters, which are few. Since strain cues operate on the feedback of possibly remote strain sources, the theory seems to complement the modern account of gene expression as being “Darwinistic” (stochastic expression of far more genes than currently needed followed by selection among resulting phenotypes) rather than “instructive” (deterministic gene expression instructed by a matched cue): strain feedback may offer a path to executing the required selection.

About Brian Cox

Ph.D. in Theoretical Physics, Monash University, 1976.
Post‐Doc appointments in Oxford, UK, and at NASA Langley: research on quantum theory of magnetism, quantum methods in fracture.
Rockwell Science Center (later Teledyne) 1981‐2015: research on nonlinear fracture mechanics, virtual tests, textile composite design, crack bridging mechanisms, energy absorbing materials, magnetically actuated biomaterials, mechanics of living systems.

Current status: gentleman scientist with active interests in strain‐cued pattern formation during development, designing composites by melding AI and human experience, and very fast fracture prediction.