Shapes in the shadow: evolutionary dynamics of morphogenesis

Modelling the evolution of novelty: a review

Evolution has been an inventive process since its inception, about 4 billion years ago. It has generated an astounding diversity of novel mechanisms and structures for adaptation to the environment, for competition and cooperation, and for organisation of the internal and external dynamics of the organism. How does this novelty come about? Evolution builds with the tools available, and on top of what it has already built - therefore, much novelty consists in repurposing old functions in a different context. In the process, the tools themselves evolve, allowing yet more novelty to arise. Despite evolutionary novelty being the most striking observable of evolution, it is not accounted for in classical evolutionary theory. Nevertheless, mathematical and computational models that illustrate mechanisms of evolutionary innovation have been developed. In the present review, we present and compare several examples of computational evo-devo models that capture two aspects of novelty: 'between-level novelty' and 'constructive novelty.' Novelty can evolve between predefined levels of organisation to dynamically transcode biological information across these levels - as occurs during development. Constructive novelty instead generates a level of biological organisation by exploiting the lower level as an informational scaffold to open a new space of possibilities - an example being the evolution of multicellularity. We propose that the field of computational evo-devo is well-poised to reveal many more exciting mechanisms for the evolution of novelty. A broader theory of evolutionary novelty may well be attainable in the near future.

Rotation of sex combs in Drosophila melanogaster requires precise and coordinated spatio-temporal dynamics from forces generated by epithelial cells

The morphogenesis of sex combs (SCs), a male trait in many species of fruit flies, is an excellent system in which to study the cell biology, genetics and evolution of a trait. In Drosophila melanogaster, where the incipient SC rotates from horizontal to a vertical position, three signal comb properties have been documented: length, final angle and shape (linearity). During SC rotation, in which many cellular processes are occurring both spatially and temporally, it is difficult to distinguish which processes are crucial for which attributes of the comb. We have used a novel approach combining simulations and experiments to uncover the spatio-temporal dynamics underlying SC rotation. Our results indicate that 1) the final SC shape is primarily controlled by the inhomogeneity of initial cell size in cells close to the immature comb, 2) the final angle is primarily controlled by later cell expansion and 3) a temporal sequence of cell expansion mitigates the malformations generally associated with longer rotated SCs. Overall, our work has linked together the morphological diversity of SCs and the cellular dynamics behind such diversity, thus providing important insights on how evolution may affect SC development via the behaviours of surrounding epithelial cells.

The sex comb (SC) is a series of modified bristles on the male forelegs of many species of fruit flies. The size, position and shape of these sex combs vary drastically across different fly species. Therefore, SCs are a model system which illustrates the interaction between evolution and organism development influencing phenotypic features. In this work, we use a combined simulation-experimental approach to study the cellular processes involved in the rotation of developing SCs in common fruit flies (D. melanogaster). Our results indicate that despite the appearance of a complicated set of motions of surrounding cells associated with SC rotation, the final SC attributes only depend on a few selected parameters. We showed that changes in the timing and extent of cell size increase in distal cells altered the extent of SC rotation and breakage. Furthermore, these changes were sufficient to account for the observed variations in SC rotation between different fly species. Thus, our computational model has given us important insights on how evolution may use various cellular processes as a means to manifest the diversity of SCs across different fly species.

On the origin of biological construction, with a focus on multicellularity

Biology is marked by a hierarchical organization: all life consists of cells; in some cases, these cells assemble into groups, such as endosymbionts or multicellular organisms; in turn, multicellular organisms sometimes assemble into yet other groups, such as primate societies or ant colonies. The construction of new organizational layers results from hierarchical evolutionary transitions, in which biological units (e.g., cells) form groups that evolve into new units of biological organization (e.g., multicellular organisms). Despite considerable advances, there is no bottom-up, dynamical account of how, starting from the solitary ancestor, the first groups originate and subsequently evolve the organizing principles that qualify them as new units. Guided by six central questions, we propose an integrative bottom-up approach for studying the dynamics underlying hierarchical evolutionary transitions, which builds on and synthesizes existing knowledge. This approach highlights the crucial role of the ecology and development of the solitary ancestor in the emergence and subsequent evolution of groups, and it stresses the paramount importance of the life cycle: only by evaluating groups in the context of their life cycle can we unravel the evolutionary trajectory of hierarchical transitions. These insights also provide a starting point for understanding the types of subsequent organizational complexity. The central research questions outlined here naturally link existing research programs on biological construction (e.g., on cooperation, multilevel selection, self-organization, and development) and thereby help integrate knowledge stemming from diverse fields of biology.

Before the endless forms: embodied model of transition from single cells to aggregates to ecosystem engineering

The emergence of complex multicellular systems and their associated developmental programs is one of the major problems of evolutionary biology. The advantages of cooperation over individuality seem well known but it is not clear yet how such increase of complexity emerged from unicellular life forms. Current multicellular systems display a complex cell-cell communication machinery, often tied to large-scale controls of body size or tissue homeostasis. Some unicellular life forms are simpler and involve groups of cells cooperating in a tissue-like fashion, as it occurs with biofilms. However, before true gene regulatory interactions were widespread and allowed for controlled changes in cell phenotypes, simple cellular colonies displaying adhesion and interacting with their environments were in place. In this context, models often ignore the physical embedding of evolving cells, thus leaving aside a key component. The potential for evolving pre-developmental patterns is a relevant issue: how far a colony of evolving cells can go? Here we study these pre-conditions for morphogenesis by using CHIMERA, a physically embodied computational model of evolving virtual organisms in a pre-Mendelian world. Starting from a population of identical, independent cells moving in a fluid, the system undergoes a series of changes, from spatial segregation, increased adhesion and the development of generalism. Eventually, a major transition occurs where a change in the flow of nutrients is triggered by a sub-population. This ecosystem engineering phenomenon leads to a subsequent separation of the ecological network into two well defined compartments. The relevance of these results for evodevo and its potential ecological triggers is discussed.

Evolutionary dynamics of RNA-like replicator systems: A bioinformatic approach to the origin of life

We review computational studies on prebiotic evolution, focusing on informatic processes in RNA-like replicator systems. In particular, we consider the following processes: the maintenance of information by replicators with and without interactions, the acquisition of information by replicators having a complex genotype-phenotype map, the generation of information by replicators having a complex genotype-phenotype-interaction map, and the storage of information by replicators serving as dedicated templates. Focusing on these informatic aspects, we review studies on quasi-species, error threshold, RNA-folding genotype-phenotype map, hypercycle, multilevel selection (including spatial self-organization, classical group selection, and compartmentalization), and the origin of DNA-like replicators. In conclusion, we pose a future question for theoretical studies on the origin of life.

A survey of application: genomics and genetic programming, a new frontier

The aim of this paper is to provide an introduction to the rapidly developing field of genetic programming (GP). Particular emphasis is placed on the application of GP to genomics. First, the basic methodology of GP is introduced. This is followed by a review of applications in the areas of gene network inference, gene expression data analysis, SNP analysis, epistasis analysis and gene annotation. Finally this paper concluded by suggesting potential avenues of possible future research on genetic programming, opportunities to extend the technique, and areas for possible practical applications.

Sociality of an agent during morphogenetic canalization: asynchronous updating with potential resonance

Canalization is a typical self-organization process leading to complementarity between parts and the whole. In the field of developmental biology, concerns about morphogenesis canalization are often framed as the French flag problem, questioning how each cell knows its own position in the whole system. Although chemical gradients have been suggested to provide positional information, there is no direct evidence that gradients are used to gain positional information. The chemical gradient hypothesis is based on the assumption that agents (e.g., cells) in a domain that locally interact with each other can generate a chemical gradient thanks to a global reference point. Instead of a chemical gradient, we here propose a model based on agents that are equipped with sociality that is based not on a global reference but rather on the ability to sense other neighboring agents, or potential resonance. The interaction among the agents with sociality, developed from undifferentiated types or tokens, is implemented using asynchronous updating automata equipped with potential resonance. We show that these automata can generate a French flag pattern that is very robust against perturbations without positional information by comparing automata with synchronous updating and asynchronous automata without potential resonance.

Evolution of networks for body plan patterning; interplay of modularity, robustness and evolvability

A major goal of evolutionary developmental biology (evo-devo) is to understand how multicellular body plans of increasing complexity have evolved, and how the corresponding developmental programs are genetically encoded. It has been repeatedly argued that key to the evolution of increased body plan complexity is the modularity of the underlying developmental gene regulatory networks (GRNs). This modularity is considered essential for network robustness and evolvability. In our opinion, these ideas, appealing as they may sound, have not been sufficiently tested. Here we use computer simulations to study the evolution of GRNs' underlying body plan patterning. We select for body plan segmentation and differentiation, as these are considered to be major innovations in metazoan evolution. To allow modular networks to evolve, we independently select for segmentation and differentiation. We study both the occurrence and relation of robustness, evolvability and modularity of evolved networks. Interestingly, we observed two distinct evolutionary strategies to evolve a segmented, differentiated body plan. In the first strategy, first segments and then differentiation domains evolve (SF strategy). In the second scenario segments and domains evolve simultaneously (SS strategy). We demonstrate that under indirect selection for robustness the SF strategy becomes dominant. In addition, as a byproduct of this larger robustness, the SF strategy is also more evolvable. Finally, using a combined functional and architectural approach, we determine network modularity. We find that while SS networks generate segments and domains in an integrated manner, SF networks use largely independent modules to produce segments and domains. Surprisingly, we find that widely used, purely architectural methods for determining network modularity completely fail to establish this higher modularity of SF networks. Finally, we observe that, as a free side effect of evolving segmentation and differentiation in combination, we obtained in-silico developmental mechanisms resembling mechanisms used in vertebrate development.

Evolutionary development of tensegrity structures

Contributions from the emerging fields of molecular genetics and evo-devo (evolutionary developmental biology) are greatly benefiting the field of evolutionary computation, initiating a promise of renewal in the traditional methodology. While direct encoding has constituted a dominant paradigm, indirect ways to encode the solutions have been reported, yet little attention has been paid to the benefits of the proposed methods to real problems. In this work, we study the biological properties that emerge by means of using indirect encodings in the context of form-finding problems. A novel indirect encoding model for artificial development has been defined and applied to an engineering structural-design problem, specifically to the discovery of tensegrity structures. This model has been compared with a direct encoding scheme. While the direct encoding performs similarly well to the proposed method, indirect-based results typically outperform the direct-based results in aspects not directly linked to the nature of the problem itself, but to the emergence of properties found in biological organisms, like organicity, generalization capacity, or modularity aspects which are highly valuable in engineering.

The evolution of robust development and homeostasis in artificial organisms

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.

During development, multicellular animals are shaped by cell proliferation, cell rearrangement, and cell death to generate an adult whose form is maintained over time. Disruption of this finely balanced state can have devastating consequences, including aging, psoriasis, and cancer. Typically, however, development is robust, so that animals achieve the same final form even when challenged by environmental damage such as wounding. To see how morphogenetic robustness arises, we have taken an in silico approach to evolve digital organisms that exhibit distinct phases of growth and homeostasis. During the homeostasis period, organisms were found to use a variety of strategies to maintain their form. Remarkably, however, all recovered from severe wounds, despite having evolved in the absence of selection pressure to do so. This ability to regenerate was most striking in an organism with a tissue-like architecture, where it was enhanced by a directional flux of cells that drives tissue turnover. This identifies a stratified architecture, like that seen in human skin and gut, as an evolutionarily accessible and robust form of tissue organisation, and suggests that wound-healing may be a general feature of evolved morphogenetic systems. Both may therefore contribute to homeostasis, wound-healing, and regeneration in real animals.

An integrated systems biology approach to understanding the rules of keratinocyte colony formation

Closely coupled in vitro and in virtuo models have been used to explore the self-organization of normal human keratinocytes (NHK). Although it can be observed experimentally, we lack the tools to explore many biological rules that govern NHK self-organization. An agent-based computational model was developed, based on rules derived from literature, which predicts the dynamic multicellular morphogenesis of NHK and of a keratinocyte cell line (HaCat cells) under varying extracellular Ca++ concentrations. The model enables in virtuo exploration of the relative importance of biological rules and was used to test hypotheses in virtuo which were subsequently examined in vitro. Results indicated that cell-cell and cell-substrate adhesions were critically important to NHK self-organization. In contrast, cell cycle length and the number of divisions that transit-amplifying cells could undergo proved non-critical to the final organization. Two further hypotheses, to explain the growth behaviour of HaCat cells, were explored in virtuo-an inability to differentiate and a differing sensitivity to extracellular calcium. In vitro experimentation provided some support for both hypotheses. For NHKs, the prediction was made that the position of stem cells would influence the pattern of cell migration post-wounding. This was then confirmed experimentally using a scratch wound model.

LinMap: visualizing complexity gradients in evolutionary landscapes

This article describes an interactive visualization tool, LinMap, for exploring the structure of complexity gradients in evolutionary landscapes. LinMap is a computationally efficient and intuitive tool for visualizing and exploring multidimensional parameter spaces. An artificial cell lineage model is presented that allows complexity to be quantified according to several different developmental and phenotypic metrics. LinMap is applied to the evolutionary landscapes generated by this model to demonstrate that different definitions of complexity produce different gradients across the same landscape; that landscapes are characterized by a phase transition between proliferating and quiescent cell lineages where both complexity and diversity are maximized; and that landscapes defined by adaptive fitness and complexity can display different topographical features.

A framework for three-dimensional simulation of morphogenesis

We present COMPUCELL3D, a software framework for three-dimensional simulation of morphogenesis in different organisms. COMPUCELL3D employs biologically relevant models for cell clustering, growth, and interaction with chemical fields. COMPUCELL3D uses design patterns for speed, efficient memory management, extensibility, and flexibility to allow an almost unlimited variety of simulations. We have verified COMPUCELL3D by building a model of growth and skeletal pattern formation in the avian (chicken) limb bud. Binaries and source code are available, along with documentation and input files for sample simulations, at http:// compucell.sourceforge.net.

A gene network model for developing cell lineages

Biological development is a remarkably complex process. A single cell, in an appropriate environment, contains sufficient information to generate a variety of differentiated cell types, whose spatial and temporal dynamics interact to form detailed morphological patterns. While several different physical and chemical processes play an important role in the development of an organism, the locus of control is the cell's gene regulatory network. We designed a dynamic recurrent gene network (DRGN) model and evaluated its ability to control the developmental trajectories of cells during embryogenesis. Three tasks were developed to evaluate the model, inspired by cell lineage specification in C. elegans, describing the variation in gene activity required for early cell diversification, combinatorial control of cell lineages, and cell lineage termination. Three corresponding sets of simulations compared performance on the tasks for different gene network sizes, demonstrating the ability of DRGNs to perform the tasks with minimal external input. The model and task definition represent a new means of linking the fundamental properties of genetic networks with the topology of the cell lineages whose development they control.

A spatial model of germinal center reactions: cellular adhesion based sorting of B cells results in efficient affinity maturation

Affinity maturation of humoral responses to T-cell-dependent antigens occurs in germinal centers (GC). In GCs antigen-specific B cells undergo rounds of somatic mutations that alter their affinity. High-affinity mutants take over GCs very soon after they appear; the replacement rate is as high as 4 per day (Radmacher et al., Immunol. Cell Biol. 76 (1998) 373). To gain more insight into this selection process, we present a spatial model of GC reactions, where B cells compete for survival signals from follicular dendritic cells (FDC). Assuming that high-affinity B cells have increased cellular adhesion to FDCs, we obtain an affinity-based sorting of B cells on the FDC. This sorting imposes a very strong selection and therefore results in a winner-takes-all behavior. By comparing our sorting model with "affinity-proportional selection models", we show that this winner-takes-all selection is in fact required to account for the fast rates at which high affinity mutants take over GCs. Another important feature of in vivo GC reactions is that they are non-mixed, i.e. GCs contain either no high-affinity cells at all or they are dominated by high-affinity cells. We here show that this all-or-none behavior can be obtained if B cells are sorted based on their affinity on the FDC surface. Affinity-proportional selection models, in contrast, always produce mixed GCs.

Origins of differentiation via phenotypic plasticity

How cell types of multicellular organisms came to be differentiated is still an open issue. Here I offer a model that posits that the origins of some cell differentiation patterns were originally passive outcomes of environmental effects. As cells' contact with the external environment was diminished, their patterns of gene expression were altered, due to changes in concentrations of externally supplied substances. Later, as multicellular growth continued, the relationships of cell layers to each other shifted, producing concentration gradients of signaling molecules. These gradients emanated both from the external cell layer toward the inside and from internal cell layers to adjacent layers. In this scenario then, differentiation arose initially as a by-product of the changing patterns of gene expression and of the complex mixtures and changing concentrations of substances passing among layers. Subsequent selection would operate to stabilize the expression patterns in those cell layers whose phenotypes provide a fitness advantage to the organism.

Computing an organism: on the interface between informatic and dynamic processes

The importance of interactions between processes at diverse space and time scales in biological information processing is a recurrent theme in the work of Michael Conrad (BioSystems, 1995: 35, 157-160; BioSystems, 1999: 52, 99-100). In this paper I present some results of explicit computational models that aim to capture and exploit some of the essential features of such multi-level processes. In the model formulation, I try to minimize the explicit definition of inter-level interactions, while providing the possibility of such interactions to develop. As often argued by Conrad inter-level interactions limit programmability. Indeed, we use an evolutionary process to derive the specific models. We study morphogenesis. We show that the interplay between cell adhesion and cell differentiation provides interesting mechanisms for morphogenesis. We also show that the interplay can both reduce and enhance small random fluctuations. We show that unequal cell cleavage in the early embryo-genesis reduces inter-individual variation of the morphemes developed from the same 'genome'. Our results suggest that, during evolution, the interplay between levels is exploited while it is at the same time reduced so as to give a certain primacy to inherited information.

Multilevel Evolution: The Fate of Duplicated Genes

Biological evolution is a multilevel process and should be studied as such. A first, important step in studying evolution in this way has been the work of Peter Schuster and co-workers on RNA evolution. For RNA the genotype-phenotype mapping can be calculated explicitly. The resulting evolutionary dynamics is dominated by neutral paths, and the potential of major change by a single point mutation.

Examining whole genomes, of which about 60 are now available, we see that gene content of genomes is changing relatively rapidly: gene duplication, gene loss and gene generation is ubiquitous. In fact, it seems that point-mutations play a relatively minor role, relative to changes in gene regulation and gene content in adaptive evolution.

Large scale micro-array studies, in which the expression of every gene can be measured simultaneously, give a first glimpse of the `division of labor´ between duplicated genes. A preliminary analysis suggests that differential expression is often the primary event which allows duplicated genes to be maintained in a genome, but alternate routes also exist, most notably on the one hand the mere need of a lot of product, and on the other hand differentiation within multi-protein complexes consisting of homologous genes.

I will discuss these results in terms of multilevel evolution. in particular in terms of information integration and the alternatives of `individual based´