Student Union Building
875 Perimeter Drive MS 4264
Moscow, ID 83844-4264
1031 N. Academic Way,
Coeur d'Alene, ID 83814
Life always finds a way. And apparently, so does simulated life.
Scientists at the University of Idaho are using a novel computer program developed on campus that optimizes complex biological systems using nature’s most potent problem solver – evolution. Based around a “genetic algorithm,” the program is currently being applied to removing ammonium perchlorate from groundwater quickly and efficiently.
“If evolution solves problems in the wild, why not apply it to practical problems in the laboratory?” says Terry Soule, professor of computer science.
Ammonium perchlorate is a substance found in groundwater across the country. It can come from rocket fuels near military bases, industrial manufacturing plants, fertilizers or simply the natural environment. Perchlorates, in general, are highly soluble, move into groundwater easily and are thought to interfere with the normal function of the human thyroid, wreaking havoc with hormones and early childhood development.
It is unclear what the exact effects of perchlorate contamination are and at what levels the contaminant becomes hazardous. Only a handful of states regulate its levels in the environment and there is no national standard. But that doesn’t mean the issue should be ignored.
“It’s like any environmental contaminant,” explains Tom Hess, professor of biological and agricultural engineering. “There’s evidence that there are health effects and you just don’t want it in the environment; especially not in your drinking water.”
To effectively neutralize the toxin, researchers are looking to nature. There are many microorganisms that degrade perchlorates into harmless byproducts. So here’s the question that University of Idaho faculty and student researchers are asking: what combination of microorganisms in which specific living conditions can get rid of the perchlorate the quickest?
This is where the genetic algorithm comes into play.
The program selects a random set of solutions by mixing the type and number of microorganisms – all taken from Idaho soil samples – as well as their living conditions by changing variables including pH levels, temperature, salinity and many others.
Scientists test the solutions to see how well they work and put the data into the genetic algorithm. The program then takes the best solutions and looks for common traits that might be important for an optimal solution. It combines these traits while introducing random mutations to create a new set of solutions. The whole process is repeated until the program evolves an effective combination.
The algorithm has many powerful applications ranging from optimizing the growth of biomass, such as the algae pictured, to developing the best nutrient combination in soil for a particular crop.
Last spring, students in Soule’s senior design class created a user-friendly computer interface for the program. Meanwhile, doctoral student Kate Kucharzyk created a faster way to test for levels of perchlorate in samples. The traditional method takes between 30 and 45 minutes for a single analysis.
“If you have 96 samples and want to do three replicates, it begins multiplying to extreme numbers,” says Kucharzyk, who is going into her fourth semester at the University of Idaho after graduating with a master’s degree in biochemistry from the Marie Curie Sklodowska University in Poland. “You’d have to run tests all day and all night.”
Instead, Kucharzyk came up with a method that uses the chemical’s excitation fluorescence to detect perchlorate levels. The process allows her to test an entire tray of 96 samples at once. With the ability to test the large number of samples required for the genetic algorithm to work, and a well-designed computer program in hand, she hopes to begin seeing results by July.
The research is funded by the Strategic Environmental Research and Development Program, a branch of the Department of Defense, and is scheduled to continue into 2010.