The Influence of Cell Type on Artificial Development

Two variants of biologically inspired cell model, namely eukaryotic (containing a nucleus) and prokaryotic(without a nucleus) are compared in this research. Experiments are designed to provide an understanding of how the evolved regulation of protein transport to and from the nucleus of the eukaryotic type cell gives rise to complex temporal dynamics that are not achievable in a prokaryoticcell. A novel system of protein movement based on the process of nucleocytoplasmic transport observed in the biological eukaryotic cell is proposed. Nucleocytoplasmic transport is considered by biologists to be one of the most important factors when determining the developmental trajectory of a cell, as it allows for additional control of transcription factors entering the nucleus, thereby regulating gene activity. Experiments contrast the ability of both cell models to generate protein patterns within the cytoplasm. Results demonstrate that the additional cell complexity of the eukaryotic does not impede the Gene Regulatory Networks control. For increasingly difficult tasks requiring precise temporal control the performance of the eukaryotic cell model outperforms the prokaryoticcell model. In addition, results demonstrate that the second level of regulation introduced by the transport process within the eukaryotic cell allows very precise control of gene activity and provides the EA with a source of heterochronic control not possible in prokaryotictype cells. Introduction Cells are the fundamental building blocks of all biological life. Two distinct groups of cell exist, eukaryotic and prokaryotic. Eukaryotic cells are distinguished by the separation of the cell into compartments, the most pronounced compartment, the nucleus contains the genome (Figure 1). The genome is decomposed into genes, which encode the blueprint for creating the organism. During the process of development gene expression results in the creation of proteins levels within the cells. Proteins within the cell direct and dictate the cell fate and define its final role within the organism. In order to enable gene expression, specialised proteins termed Transcription Factors (TF) must bind the gene cisregulation sites. In response to TF binding, the gene expresCytoplasm Nucleoplasm Genome Outer Cell Wall Nuclear Pore Complex Figure 1: Architecture of the eukaryotic cell model. The nucleus is composed of the Nucleoplasm, Genome, and Nuclear Pore Complex (NPC) sion rate is regulated, influencing protein levels within the cell. The presence of a nucleus within eukaryotic cells (Figure 1) restricts direct entry of transcription factors to the nucleus. Transport across into the nucleus through the Nuclear Pore Complex (NPC) is enabled by specialised chaperons proteins binding to TF proteins. The chaperon proteins are subdivided into two categories importins which enable import to the nucleus, and exportins which enable export from nucleus to cytoplasm across the NPC. Importins bind specific sites on the TF, named Nuclear Localisation Sequence (NLS). Export from the nucleus is enabled by the binding of exportin to Nuclear Export Sequence (NES) associated with the TF. This network of interacting genes and proteins is known as the Gene Regulatory Network (GRN). Modifications to the relative timing of developmental events are termed heterochronic. Lee and Hannink (2003) report that the control of protein entry to the nucleus provides a powerful mechanism for the temporal regulation of gene expression. West-Eberhard (2003) highlights the importance of heterochronic control as a source of phenotypic novelty “if I could control the time of gene action I could cause the fertilised snail egg to develop in an elephant” Heterochronic change is not a developmental process but rather an evolutionary process (West-Eberhard, 2003). Modifications to the timing of events are heritable and must be somehow linked to the genetic encoding. Artificial developmental systems have been introduced as a technique aimed at increasing the scalability of Evolutionary Algorithms (EA) (Haddow and Hoye, 2007). These ECAL General Track