The Digital Evolution Laboratory at Michigan State University
Contact
AboutPeopleProjectsPublicationsLinks
>> Home > Projects

A sampling of ongoing research projects in the Devolab

Trace of evolved, complex lineage

The Evolutionary Origin of Complex Features.

From the time Darwin first proposed his theory of evolution by natural selection, its critics have raised the question of how new complex traits can arise. Using digital organisms, however, we can have full information about a population as it evolves and trace back the linage of any individual evolved organism (as shown in the image to the right). From this, we can study the effects of each mutation along this line of descent, and determine the exact mechanisms that drive these evolutionary events. [MORE]
Researchers: Lenski, Ofria, and Pennock (with Chris Adami, Keck Graduate Institue)
Contact: ofria@msu.edu

Teaching Evolution and the Nature of Science using Digital Organisms

The Avida-ED Project is developing and classroom-testing an educational version of Avida together with supporting curricular materials for use in undergraduate biology lab courses. Using the evolutionary principles instantiated in the digital environment, students can learn about complex systems and their emergent properties. Guided exercises built around such inquiry-based experiments can also help students learn about the nature of scientific evidence and reasoning. We will assess the effectiveness of this new technology in the classroom and disseminate the software and materials nationally. [MORE]
Researchers: Pennock, Clune, Nanlohy, Ofria, and Bryson
Contact: pennock5@msu.edu

Survival of the Flattest

When organisms have to evolve under high mutation pressure, their evolutionary dynamics is substantially different from that of organisms evolving under low mutation pressure, and some of the high-mutation-rate effects can appear paradoxical at first glance. In Avida, we we subjected digital organisms to extremely high mutation rates and then examined the effect on the genetic architecture of the organisms. [MORE]
Researchers: Ofria and Lenski (with Claus Wilke, Keck Graduate Institue)
Contact: ofria@msu.edu

Phylogentic visualization from an ecosystem

Evolution of Digital Ecosystems

Most Avida experiments to date have been performed using only a single niche population -- the organisms have unlimited resources to work with and thus multiple species cannot co-exist in the long-term. However, we have started research in Avida where we limit the available resources and from this we witness the spontaneous emergence of simple ecological communities. [MORE]
Researchers: Ofria and Lenski (with Tim Cooper, University of Auckland)
Contact: ofria@msu.edu

Phylogenetic Reconstruction: Discovering the tree of life

A major challenge in evolutionary biology is how to determine the historical relationships among organisms (their "phylogenetic tree"). In the biological world, it is difficult to gauge the accuracy of a reconstructed tree (and thus its associated reconstruction algorithm) as most of the information we need is lost to history or to the impossibility of perfect measurement. However, in a transparent system such as Avida, we can fully record this information during the original evolution, and then compare the output of each algorithm to the true tree to isolate the situations in which they perform best. This can allow for advances in areas such as disease tracking, our understanding of gene function, and pharmaceutical drug design.
Researchers: Torng, Schmidt, Ofria, Jin, Hang, Huang, and Rupp
Contact: torng@msu.edu

Evolving Intelligence (EI)

As an alternative to traditional artificial intelligence (AI) research which aims to design intelligent systems from the top down, we think that a more promising approach is to evolve intelligence from the bottom up. Using evolving digital organisms we are exploring patterns in the emergence of simple intelligent behavior. [MORE]
Researchers: Pennock, Ofria, Lenski, Clune, Elsberry, and Grabowski.
Contact: pennock5@msu.edu

The Evolution of Sex

The origins and maintenance of sex and recombination are poorly understood, even though sexual organisms are found virtually everywhere in the biological world. We are exploring many of the proposed theories about the evolution of sex by testing them in Avida. We have started this line of research with a study of the Muller's ratchet hypothesis, which states that recombination can help restore unmutated genomes when a mutation rate would otherwise be too high for the population to survive. We have found that sex does indeed contribute to survival under strong genetic drift for a narrow range of conditions [Misevic04]. In our current project, we are studying hypotheses that relate the importance of sex to changing environmental conditions.
Researchers: Misevic, Lenski and Ofria
Contact: misevicd@msu.edu

The Evolution of Genetic Architecture

Modules have been studied as the units of biological organization in many diverse contexts such as development, metabolic networks, and genome organization. However, the origins and causes of modularity have been explored much less then its consequences. Using Avida, we qualify the influence of mode of reproduction on modularity and epistasis in digital organisms.
Researchers: Misevic, Lenski, Ofria
Contact: misevicd@msu.edu

Group Selection

In a colorful example of group selection, Craig and Muir [Craig96] showed that selection on whole groups of chickens living in the same cage resulted in a dramatic increase in individual productivity and survival. Using digital organisms instead of poultry (luckily!), we explore how selection on the level of a group can interact with individual selection and potentially explain the evolution of sex or high mutation rates, traits often thought to be detrimental for the individual, but beneficial for the group.
Researchers: Clune, Misevic, Pennock, Ofria and Lenski
Contact: jclune@msu.edu

The Evolution of Mutation Rates

The rate at which genomes are mutated is a crucial driving force of evolution but is also under selection itself. We combine analytical modeling and Avida experiments to describe the interaction of costs and benefits of high mutation rates, and predict the course of mutation rate evolution in simple and complex environments.
Researchers: Misevic, Clune, Ofria and Lenski
Contact: misevicd@msu.edu

Philosophical Issues in Evolutionary Computation and Modeling

The ability to evolve artificial lifeforms allows one to raise and address a variety of interesting conceptual questions in the philosophy of biology. How should one define life and recognize something as a lifeform? What ethical issues need to be considered in such research? In what ways are digital organisms simulations and in what ways are they instances of biological systems? How do different sorts of models affect their evidential value? How can evolutionary theory be generalized beyond the simple organic case? Artificial lifeforms allow empirical investigations of conceptual possibilities that are otherwise difficult to test. Current projects are investigating the working of group selection and the evolution of altruism.
Researchers: Pennock and Clune
Contact: pennock5@msu.edu

Meta-EC: Configuring Evolutionary Computation via Evolutionary Computation

One of the major promises of evolutionary computation (EC) is having computers solve difficult problems with minimal human intervention. In reality, however, getting EC to provide solutions to problems requires an extreme amount of arcane knowledge about how to choose a satisfactory setup of evolutionary parameters among the enormous number of possible configurations. This is due to the fact that what constitutes a good EC setup -- which genetic operators to use and with what frequency -- changes from problem to problem such that a poor configuration will fail to yield a valuable result. Given that EC itself is good at finding satisfactory solutions amongst large multi-dimensional search spaces, it makes sense to try to find good EC setups using EC. We explore the viability of this strategy, which we call "meta-EC" and preliminarily results indicate that it is a good means of automating the process of selecting parameter settings for EC. If successful, meta-EC research could help EC deliver its original promise.
Researchers: Clune, Goings, Goodman and Punch
Contact: jclune@msu.edu

Evolved Wire-Frame Structure

Wire-Frame Organisms - Evolution in a virtual 3-Dimensional world.

Evolving structures in a virtual 3-D space that has physical attributes such as gravity and friction allows us to test evolution in a different kind of environment than in the Avida system. Can "useful" structures evolve in this environment? Can these structures be translated into the real world? Can we learn more about the process of evolution if we work with organisms that have (virtual) physical bodies?
Researchers: Stredwick, Covert, and Ofria
Contact: stredwic@msu.edu

Last Updated Fri Jul 18, 2008