The US Environmental Protection Agency last week announced that it had selected Amersham Biosciences’s CodeLink microarrays to perform gene-expression experiments to discover biomarkers for aquatic toxicity.
Financial details of the agreement were not disclosed.
The deal, however, served to launch a new custom design and manufacturing service for Amersham, which is slated to merge with General Electric.
Separately, in response to questions raised over CodeLink’s future under the $9.5 billion merger, Alexandra Morris, a company spokeswoman, said, “our customers will be able to continue over time to use the CodeLink products.”
EPA will use the CodeLink service to manufacture microarrays based on gene sequences from the fathead minnow (Pimephales promelas), the vertebrate aquatic toxicological model organism. EPA researchers will use the microarrays to determine biological indicators for identifying contaminants in aquatic environments.
David Lattier, one of the EPA principal investigators on the project, told BioArray News that the group selected the CodeLink platform because of its technology – specifically, the method the company uses to deposit DNA onto a three-dimensional substrate.
“Their method of spotting allows all of the targets on a microarray to be available for hybridization, and that was critical for reproducibility,” Lattier said. “This spotting is very critical, and the other platforms just don’t compare.”
Lattier said that the researchers will be using low-density microarrays in their research and will be dealing with an organism whose genome has not been fully sequenced. The scientists have identified 150 genes, specific for fathead minnows, and will be attempting to create arrays in an “intelligent design” approach, Lattier said. The genes will be validated through exposure to estrogenic compounds.
“One of things we are trying to do is use low-density microarrays with specific genes that will be regulated, either up or down, by environmental stressors,” Lattier said. With high-density microarray platforms, “signal-to-noise ratios tend to get into the way. Also, recent publications suggest that there are gene-expression neighborhoods, or eukaryotic operons, if you will, where a lot of genes are influenced [toward] transcription based on position and not necessarily as a result of specific induction.”
As the research moves forward, more genes will be added to the microarrays.
The way the CodeLink platform is set up, we can continue to add sequences with very little problem,” he said. “It will be fairly trivial, actually.”
The company is now in the process of creating a memorandum of understanding with the EPA, and the DOE’s Joint Genome Institute in Walnut Creek, Calif., to facilitate the research process.
“The amount of effort other organizations have on the high throughput sequencing end, other organizations dwarf what EPA is doing,” said Greg Toth a program manager for computational toxicology at the EPA’s National Exposure Research Laboratory. “Because they are a federal organization, there is an interest in partnering with other federal organizations to expand what we can now do in ecological risk assessments. We are going to give them cDNA libraries, and they will sequence those.
“The ability to enhance what is on the arrays will increase incredibly rapidly,” he said.
The research will examine types of stress that humans and wildlife are exposed to, Toth said.
“It’s a real black box and an enormous amount of work with arrays has gone on to try to understand what is going on with adverse affects in organisms and toxicity pathways,” he said. “Our work here is to identify the components that are most important in environmental exposures and what that leads to is to make sense out of mixtures. We are poised to look back to sources of exposures, hopefully to use the data we get from the arrays, to identify what the critical components are of the mixtures, and we can look downstream and hopefully link the indicators of exposure we have to indicators of adverse affect, and ultimately to outcome.”
The research will first come up with an array, and then use it to identify changes in expression that are unique for a variety of stressors.
“In the past, with the molecular tools that we had, we could look at one stressor at a time,” said Toth. “Now in two or three years, we should be able to ID the first signatures of exposures to perhaps 22 to 30 different stressors. With the array, we are looking at having between  and 400 genes, and [with] the interaction with JGI ... [which] will start in 2004, we can have somewhere in the low thousands of genes. Once we have gotten to that point, it’s a matter of doing exposures to the organism and then doing the analysis.”