From Santa Fe Institute Events Wiki
My background is in both biology and computer science. I have graduated from the University of Gdansk (Poland) in 1997, where I was involved in a number of projects on bacterial and viral molecular genetics. I have then spent a year as a Fulbright Visiting Researcher in the Salk Institute for Biological Studies in San Diego, CA (where I worked on gene therapy in the brain using AAV vectors). After that, I went back to Poland to defend my PhD thesis in molecular genetics at the University of Gdansk. Two months later, I started my first postdoc at the Hebrew University of Jerusalem (Israel), again in the hope of developing ways to do gene therapy in the brain. After I gave up learning Hebrew for hopeless (hey, I can still say 'hello!'...and... that's about it) I suddenly started to have lots of free time between the experiments. So I brushed up on my programming skills and did some bioinformatics on the side... This has got me so hooked up that for the next postdoc I went to Valencia), Spain) (at least I was consistent with climatic zones) to work on molecular evolution of HCV at the Cavanilles Institute. Valencia is a great city for skating and this skill was crucial when going from one campus to another, because I decided to attend classes in computer science and maths at the University of Valencia. My work on phylogenetics continued after I returned to Poland to work at the Institute of Oceanology, but again slowly the main focus of my research shifted. The main main research interest now is to develop a biologically realistic artificial life model to study the evolution of complexity at various levels: the level of the gene regulating network that governs the behavior of single cells (evolution of GRNs), the level of multicellular bodies (co-evolution of mind and body) and the level of communities of organisms (evolution of communication and co-operation in a multi-agent system).
A lot more about my research project than you would care to know
In several recently published papers (Joachimczak & Wróbel 2008ab, 2009) I have presented preliminary results on an artificial evolution system in which complex mutlicellular bodies develop from single cells. In such a system, simulated cells divide and differentiate, and the phenotype of the cells/bodies is coded indirectly in the artificial genome. Variation in the population is introduced by imperfect copying of the genomes during reproduction. I would like to extend this system now to allow the simulated biological-like objects (animats) to evolve in an artificial environment with limited resources. Survival and reproduction will depend on the ability to use these resources (chemical substances present in the environment), to find mates, and to avoid obstacles and predators. Building such a system requires careful design in which the biological realism has to be balanced with computational limitations.
The physics in such an artificial system has to be simplified. The environment that has been implemented thus far consists of a viscous liquid in which chemical substances (signaling presence of food, mates or predators) will diffuse. Flagellum-like actuators can be used to propel the simulated biological-like objects. Simplified physics governs also the mutlicelluar development. At the level of single cells, the products of the genome (so called “internal products”) have affinity for promoters and control the expression of other genes. Special class of products (morphogens) may diffuse from the source cell (forming a gradient of concentration). These substances allow for cell-to-cell communication: the genome can also code for receptors that have affinity to morphogens. The interaction between differentially expressed receptors and morphogens allows cells to direct division or differentiation. The papers published thus far describe the details of the model and some preliminary results, showing that it allows to evolve regulatory networks controlling development of non-trivial 3D shapes consisting of hundreds of spherical cells.
To be relevant for biological research, a model of artificial life should include interaction between the organisms and their environment (the physical features of the environment: the particular laws of physics, such as gravity, friction/viscosity, diffusion of substances...). In "How the Body Shapes the Way We Think" (2006, MIT Press) Pfeifer and Bongard argue that the environment shapes the biological organisms at three temporal scales: (i) the phylogenetic, or evolutionary scale (the information in the genome may be interpreted as a representation of information about the environment to which the ancestors of the organism have been exposed), (ii) the ontogenetic scale (embryogenesis, morphogenesis, development), and (iii) the here-and-now scale of the moment-to-moment interaction of the organism with its environment, as it obtains resources in order to reproduce (with possibly additional elements that it involves, for example, the search for a mate).
I aim to build an artificial life system in which all these temporal frames are present: evolution, morphogenesis and behavior. The research platform thus created will team up ideas, concepts, and modeling frameworks on two levels: indirect genotype-phenotype relationship through the application of GRNs, and genome-phenotype-physics interaction for shaping mutlicellular bodies/communities. I hope that the development of this research platform will be a very important contribution to the field of artificial life and theoretical biology. I would like to use this platform to explore important questions in the fields of evolutionary biology, developmental biology (and at their interface, evo-devo), systems biology, and also cognitive science/theoretical neurobiology. Such questions include, firstly, the investigation of robustness of developed networks, and of the statistical properties of the biologically-inspired networks. At the level of cell communities and multicellular bodies, important questions include the evolution of cooperation between the cells, and evolution of symmetry breaking and patterning mechanisms during development of multicellular bodies (complexification/modularity in development and form, the emergence of developmental symmetry). Finally, at the level of the emerging ecosystem of animats, interesting questions include the evolution of complex behavior (including communication) in a system in which co-evolution of mind and body is possible.
I have used several programming languages (C, C++, Java, Perl, Python) and environments (Matlab, R, etc).
What I hope to learn in Santa Fe & possible projects
I hope that the participation in the Complex Systems Summer School 2010 will allow me to learn the modern tools of complex system analysis that I could then include in the research platform I am building. This tools will be then used to analyze the properties of artificial GRNs, artificial neural networks that will regulate the animat behavior, and the properties of the ecosystem of co-evolving animats. Although a complete system is not build (and no, we do not sell it yet...), the software already develop can produce a huge number of networks that were evolved to do a specific task. For example, it is possible to use a genetic algorithm to obtain networks able to generate/process signals (for example, generate oscillations or work as low-pass filters). Any help will be appreciated to develop ways to analyse what are the properties of these networks!