Home Brief CV Projects Publications Teaching Contact

Some Recent Projects -- Bioinformatics


GAMI: genetic algorithms for motif inference
Synopsis: In this project, we develop computational tools to infer regulatory regions in noncoding DNA, based on conservation over evolutionary time, and including biochemical data from the ENCODE project.
Current and recent collaborators: Leif Oxburgh, Carol Kim, Jim Coffman
Current staff and students: Craig Lessard, Sam McFarland, Brenna Gardner, Naomi Heuer, Amanda Castonguay, Ben Congdon-Jones
  • With Carol Kim's lab at the University of Maine, we have computationally identified regulatory elements for CFTR, the gene responsible for cystic fibrosis, and completed preliminary biological validations of these.
  • With Jim Coffman's lab at the Mount Desert Island Biological Lab, we have computationally identified regulatory regions for FOXJ1, a gene important in cilia development, and completed preliminary biological validations of these.
  • With Leif Oxburgh's lab at the Maine Medical Center Research Institute, we have computationally identified regulatory regions for PDGFRb, a gene critical to kidney development, and completed preliminary biological validations of these.

GPS: Genome Pattern Scan
Synopsis: In this project, we develop computational tools that identify genes that might be influenced by specific regulatory factors (specified by the user).
Collaborators: Leif Oxburgh, Mary Ann Asson-Batres
Staff and students: Sam McFarland, Craig Lessard, Jen Fetting
  • With Leif Oxburgh's lab at the Maine Medical Center Research Institute, we are working to identify genes that might be regulated by FOXD1.
  • With Mary Ann Asson-Batres of Tennessee State University and Doug Sawyer of the Maine Medical Center Research Institute, we are working to identify genes that might be regulated by the presence of Vitamin A.

Gaphyl: genetic algorithms for phylogenetics
Synopsis: Find the most plausible evolutionary relationships among species
Staff and students: Emily Greenfest, Josh Ladieu, Kevin Septor, Lunn Sawyer


Some Recent Projects -- AI and Games


REALM: A Rule-Based Evolutionary Computation Agent that Learns to Play Mario
Winner of the 2010 Mario Learning Track and the 2010 Mario Gameplay Track at the Computational Intelligence and Games Conference, Copenhagen Denmark, August 2010
Videos of the agent
Synopsis: Using an evolutionary computation approach to learn to play Mario.
Masters project by Slawomir Bojarski

Learning to Play Unreal
Winner of the 2009 Unreal competition at the Congress on Evolutionary Computation, Trondheim Norway, May 2009
Synposis: Development of Chimera, an evolutionary approach to learning rule-based systems and Agent Smith, an agent to play Unreal. A hand-coded version of Agent Smith won the CEC 2009 Unreal Tournament competition.
Independent study project by Ryan Small


RAMP: A Rule-Based Agent for Ms. PacMan
Winner of the 2008 Ms. PacMan competition at the World Congress on Computational Intelligence, Hong Kong, May 2008
Synopsis: Artificial intelligence approach to game playing.
Masters project by Alan Fitzgerald
Contributors: Peter Kemeraitis
Base code and competetion run by Simon Lucas