Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Filters applied. Clear all
, 4 (3), e1000030

The Evolution of Robust Development and Homeostasis in Artificial Organisms


The Evolution of Robust Development and Homeostasis in Artificial Organisms

David Basanta et al. PLoS Comput Biol.


During embryogenesis, multicellular animals are shaped via cell proliferation, cell rearrangement, and apoptosis. At the end of development, tissue architecture is then maintained through balanced rates of cell proliferation and loss. Here, we take an in silico approach to look for generic systems features of morphogenesis in multicellular animals that arise as a consequence of the evolution of development. Using artificial evolution, we evolved cellular automata-based digital organisms that have distinct embryonic and homeostatic phases of development. Although these evolved organisms use a variety of strategies to maintain their form over time, organisms of different types were all found to rapidly recover from environmental damage in the form of wounds. This regenerative response was most robust in an organism with a stratified tissue-like architecture. An evolutionary analysis revealed that evolution itself contributed to the ability of this organism to maintain its form in the face of genetic and environmental perturbation, confirming the results of previous studies. In addition, the exceptional robustness of this organism to surface injury was found to result from an upward flux of cells, driven in part by cell divisions with a stable niche at the tissue base. Given the general nature of the model, our results lead us to suggest that many of the robust systems properties observed in real organisms, including scar-free wound-healing in well-protected embryos and the layered tissue architecture of regenerating epithelial tissues, may be by-products of the evolution of morphogenesis, rather than the direct result of selection.

Conflict of interest statement

The authors have declared that no competing interests exist.


Figure 1
Figure 1. Evolving multicellular digital organisms.
(A) The genomes of digital organisms used in this study are made up of 100 rules. 4 numbers define each rule. Digit 1 determines whether the rule is contingent on space (local neighborhood in 3D space) or time (number of divisions). Digits 2 and 3 define the minimum/maximum range of action of each rule (number of local neighbours or interval), and Digit 4 defines the type of action or anti-action to be implemented (to clone a new cell in an adjacent location in the Moore neighbourhood, to move to a neighbouring cell, or to die). An example of a rule is given in full. (B) A genetic algorithm directs organism evolution. (C) The genome of each organism guides its development from a single cell. After allowing for an initial period of growth, organisms are selected that exhibit morphological homeostasis for a period of 100 time-steps (at t = 50, t = 100 and t = 150). Selection is also used to favour organisms that have a surface volume ratio of 0.8 which do not cross the boundary of the 50 by 50 matrix, and to select against fragmented organisms.
Figure 2
Figure 2. Static and dynamic homeostasis.
The pictures are stills of the development of 3 organisms at time steps 50, 100 and 150. Cell age is depicted using a colour key. The organisms exemplify the different types of homeostatic behaviour observed. (A) Organism #11 is static and maintains its form by ossifying, limiting the rate of cell birth, death and movement. (B) Organism #17 is dynamic and maintains an evenly balanced but high rate of cell birth and cell death. (C) In organism #18, dynamic and static regions coexist, so that cells are born at its base and die some time later in the upper regions. This generates a visible gradient of cell ages from lower (old-red) to higher (young-blue) planes. (D–F) Graphs show the impact of mutating individual genes in the genome of organisms #11(D), #17 (E) and #18 (F) on homeostasis. Genes with the highest impact on homeostasis (genes #14 and #94 in organism #17 and gene #80 in organism #18) all encode regulators of cell death (see Videos S2B and S2C).The impact of each mutation was calculated using a combination of 2-point correlation and lineal path analysis between time-steps 50, 100 and 150.
Figure 3
Figure 3. The ability of homeostatic organisms (A) #11, (B) #17, and (C) #18 to withstand environmental perturbation was tested by inducing wounds at time step 100.
The recovery process was then followed over time.
Figure 4
Figure 4. The development of a robust stratified tissue architecture.
A series of analyses were used to examine regional cellular behaviour in organism #18. (A) The rates of cell proliferation and (B) the direction of cell movement in organism #18 were calculated layer by layer and are depicted as a bar chart. Organism #18 was then wounded by removal of X cells from (C) the top part or (D) the bottom part, and the response followed over time as indicated. Ancestors of organism #18 were then traced back in evolutionary time using a BLAST-type algorithm. The earliest ancestor with a similar genotype and phenotype was identified in generation 3, after two rounds of selection (see Table S1). (E) This organism was wounded as above, but is unable to heal.
Figure 5
Figure 5. The recovery of organism #18 at generation 30 is shown following wounding at time-step 100 after the removal of a plane of cells at 3 different heights.
The colours denote the action being implemented: red denotes cell death and yellow a cell cloning.
Figure 6
Figure 6. Quantitative analysis of the evolved wound-healing response.
(A, B) For this analysis, a wound was generated by removing a slice with thickness equal to five voxels from a central layer in organism #18 at time-step 100. (A) Healing is almost complete after 30 time-steps in the evolved organism at generation 30, whilst the wound remains open at the equivalent time-step in the ancestral organism. In both cases, gene 67 is activated at the wound margin following wounding (highlighted in light blue). (C) A genetic analysis of wound healing in organism #18 was carried out. Each gene was eliminated in turn and the average impact on healing determined using a combination of 2-point correlation and lineal path analysis. 3 lateral wounds (removal of planes at different positions in the X axis at time-step 100) were generated, the average and standard deviation calculated, and the graph normalized to ensure that defects in homeostasis do not confound the analysis. Gene 67 (if (west in interval [7-7] then clone in dir (-1,0,0)), has the greatest relative impact. (D) The proportion of cells in which gene 67 is active is depicted for each time-step during the wounding experiment shown in (A) (see Video S5A). The gene is activated by the morphological changes that accompany wounding (see Video S9) and plays a role in wound-healing (see Video S10).
Figure 7
Figure 7. The impact of single gene mutations on homeostasis in organism #18 at generations 3 (red) and 30 (green) are compared.
The average impact of a mutation on homeostasis was 0.74 +/− 2.99 at generation 3, and 0.09, +/− 0.12 at generation 30.
Figure 8
Figure 8. The recovery of organism #18 is depicted following wounding in the lateral plane (A) at generations 3 and 30 (see Videos S11 and S12 for the organism at generation 3 with and without gene 67) and (B) at generations 3, 10, 20 and 30 in the evolutionary process.
Wounds (5 voxels in width at position 25) were induced at time-step 100 and the recovery of form followed using a 2 point correlation and lineal path analysis over time. The graphs were normalized for wound size.
Figure 9
Figure 9. The recovery of organism #18 at (A) generation 30 and (B) generation 3 is shown following wounding at various positions in the Y axis at time-step 100 (see Video S1F) and in (B).
The impact and recovery were followed using a 2 point correlation and lineal path analysis over time. Graphs were normalized for wound size.

Similar articles

See all similar articles

Cited by 6 PubMed Central articles

See all "Cited by" articles


    1. Wolpert L. Principles of Development. In: Wolpert L, editor. Principles of Development: Oxford University Press; 2007.
    1. Wagner A. Robustness, evolvability, and neutrality. FEBS Lett. 2005;579:1772–1778. - PubMed
    1. Lucchetta EM, Lee JH, Fu LA, Patel NH, Ismagilov RF. Dynamics of Drosophila embryonic patterning network perturbed in space and time using microfluidics. Nature. 2005;434:1134–1138. - PMC - PubMed
    1. Kollmann M, Lovdok L, Bartholome K, Timmer J, Sourjik V. Design principles of a bacterial signalling network. Nature. 2005;438:504–507. - PubMed
    1. Ciliberti S, Martin OC, Wagner A. Robustness can evolve gradually in complex regulatory gene networks with varying topology. PLoS Comput Biol. 2007;3:e15. - PubMed

Publication types