Parameterisation of developmental networks to understand periodic patterning

Lead Research Organisation: Heriot-Watt University
Department Name: S of Mathematical and Computer Sciences


Advances in genetics have given us a good idea of how many genes we have. From this we get an idea of how many components act to make and maintain our bodies. However, it is clear that genes do not act in isolation, but act with others to form a network. The operation of the network as a whole, rather than that of an individual gene, gives tissues their shape and cells their identity.

Even before genes were isolated and sequenced, ideas were formed concerning the nature of network interactions that might operate to drive the development of form in the embryo. Almost 60 years ago Alan Turing suggested that many examples of repeated strucutres in nature, such as the alternating leaves on a plant or the spacing of hairs in the skin, could be explained by a simple set of interating substances, which he called morphogens. He realised that if two morphogens have opposing properties, but operate over different ranges, then this system can produce ordered patterns out of almost nothing. Since this seminal work theoreticians have applied this insight to a wide range of biological patterns, such as spot and stripe markings on animal skin, the colour patterns on seashells, internal organs such as the trachea and even the distinction between the left and right sides of the body. Theory is beginning to meet experimental work in defining what Turing's postulated 'morphogens' might be for specific systems, but so far studies in this area have been limited to labelling this or that molecule as a morphogen. To move this field forward we need (i) a definition of the important morphogens that act in a specific tissue, (ii) a mathematical model that can describe the operation of these molecules and (iii) a set of measurements for the important characteristics of these molecules, such as the speed with which they move through tissues and the their molecular lifespan. We will perform a bottom-up integrated mathematical and experimental study of feather patterning in embryonic chicken skin to fully test Turing's idea for the first time.

Recent genetic advances have allowed us to define the key molecules required to produce a pattern in the chicken skin, giving us the key components we need to piece together an understanding of this system. Embryonic skin develops well in culture, allowing us to observe and manipulate the patterning process, and provides a large amount of material for experimentation. This is therefore the ideal system in which to test our mature ideas about how higher level networks of genes operate to create ordered patterns in nature.

Technical Summary

Periodic patterns are a fundamental feature of vertebrate anatomical organisation. Such patterns, arising rapidly in the embryo, have long been suspected to involve the action of a system involving coupled short range activatory and long range inhibitory processes. Five decades of theory have clearly shown that a system of such interactions can yield self-organisation and the production of organic-looking patterns, at least on a computer screen. While developmental biology, genetics and molecular biology have begun to explore what the components of such systems might be, this has not yet advanced to the stage at which we can properly test or understand its operation. To move forward a quantitative molecular approach is needed to provide the hard data to test the theory and illuminate its operation in a defined system.

We will build on our recent work in avian genetics and developmental biology to establish a structural network of signalling interactions that patterns the developing chicken skin. We will produce a mathematical model that describes these interactions and which takes into account realities in biological processees, such as defining the production of proteins in terms of transcription and translation. We will then measure the relevant molecular parameters, such as diffusion and decay rates of proteins in the system, to input into the model. The output of this simulation will inform us as to how well it can explain the size and spacing of pattern elements in the skin, as well as the timescale of pattern emergence. The simulations will be used to predict the effects of alterations to molecular parameters and signal intensities, and these predictions will then be tested experimentally to refine and evaluate the model's predictive power. This iterative experimental/modelling approach will then be expanded to incorporate more entities into the network, if necessary, and to explore the malleability and robustness of the system.

Planned Impact

This is primarily a fundamental science project that aims to produce insights into the mechanisms of embryonic development and pattern formation. The immediate impact will chiefly be upon the academic beneficiaries and the general public, with industrial benefits arising from our work on the genetics of feather control, which is clearly documented as a very beneficial trait for egg and meat production, and animal welfare, in hot conditions (either chronic or acute). Our explorations of the recessive scaleless mutant have resulted in a genetic test for the mutation which will be reported along with our findings in development, signalling and pattern formation. This test can be used for rapid introgression of this trait into any desired line, particularly anticipated to be useful for the broiler chicken industry.

The public impact of this project upon the public will be enhanced by the 'real-world' nature of the model system studied; the distribution of hair of different types across the body. This model is easily appreciated by a lay audience and, combined with the striking images produced in studies of spatial patterning, this area is ideal for public engagement activities. In this regard the use of cultured skin and computer simulation where appropriate in the project, rather than Protected Animals, will aid public engagement and acceptance.

Academic beneficiaries will gain new insights into the basis of vertebrate development and the way in which computational tools can be used in conjunction with wet experiments to attain a deeper understanding of biological phenomena. In addition, the specific signalling pathways under study are relevant to the development of many vertebrate organs and will stimulate studies regarding the integration of these signals in a range of tissues.


10 25 50
Description We have developed a mathematical/computational model to describe the developmental pattern forming process that leads to periodically arranged feather buds on the skins of chickens (along with other birds). The mathematical model incorporates some of the major components involved in patterning: the cell populations of both the epithelium and mesenchymal skin layers and key signalling molecules FGF and BMP. In particular, the model has been formulated according to a minimal set of interactions between these components as observed by our experimental collaborators: in other words, we have proposed the model according to a set of interactions based on experimental observations, rather than intangible interactions for which no evidence has been found.

This basic model has been shown (through a combination of both analysis and numerical simulation) to be capable of patterning the skin into spatially periodic structures in a manner closely resembling the real-life formation of feather placodes. A key finding in this work is that a classic activator-inhibitor mechanism is not necessary for patterning, as suggested in many previous modelling studies. Instead, here the pattern formation process is driven by cellular movement in response to signalling molecules. Our ongoing work is to explore the model's capability in explaining various experimental findings, and therefore test the degree to which the current model can describe patterning.
Exploitation Route Our findings are stressing the key contribution of cellular movement to pattern formation. Our current model is based on a classical cell movement model of continuous type, i.e. one in which cell populations are represented by continuous cell density distributions. Optimal utilisation of the experimental data, however, may be better served via a discrete cell based approach, and this is a key area for future model development.

At a broader level, the continuing development of computational/mathematical models that capture the basic interactions observed experimentally should allow them to slot naturally into the experimental workflow by offering a quick and cheap method for testing the viability of different mechanisms, thereby providing a pathway for replacement/reduction of experimentation.
Sectors Digital/Communication/Information Technologies (including Software),Manufacturing, including Industrial Biotechology,Pharmaceuticals and Medical Biotechnology

Title Computer model for patterning 
Description We are developing a computer model to describe patterning development. The ultimate aim of this is to provide predictions and reduce/replacement experimentation (3Rs) 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Not at the present time: model is still in developmental phase 
Description Dr Denis Headon 
Organisation University of Edinburgh
Department The Roslin Institute
Country United Kingdom 
Sector Academic/University 
PI Contribution We are developing models to help understand the biological systems being studied by the Headon group.
Collaborator Contribution The Headon group are providing us with the biological data and expertise necessary to formulate our models.
Impact Multidisciplinary (mathematical-biological). We are currently writing a number of papers based on the work in this collaboration.
Start Year 2009