Lawrence Berkeley National Laboratory
At a Glance
Name: Gary Andersen
Title: Scientist, Center for Environmental Biotechnology, Lawrence Berkeley National Laboratory
Professional Background: 2003 — present, group leader, molecular and microbial ecology, Earth Sciences Division, Lawrence Berkeley National Laboratory; 1997 — 2002, senior biomedical scientist, Biodefense Division, Lawrence Livermore National Laboratory.
Education: 1993 — PhD, University of California, Berkeley, plant pathology; 1982 — MS, University of California, Berkeley, plant pathology; 1978 — BA, Northwestern University, biology.
Representatives of the US Army Corps of Engineers this week announced that they have succeeded in pumping most of New Orleans dry, days ahead of an Oct. 8 deadline they had set for themselves.
However, the floodwater, described in various media outlets as being contaminated by raw sewage and gasoline, has left behind a crusty residue of bacteria that can still threaten inhabitants as they resettle the fetid city.
To test the residue, researchers at Lawrence Berkeley National Laboratory in Berkeley, Calif., are recommending the use of a pathogen-detecting microarray they have developed, known as the phylochip.
Using the 500,000-probe array, users can detect the presence of up to 9,000 bacterial agents, according to a statement from LBNL. To learn about the new offering and how it can be applied to real-life situations, BioArray News corresponded with Gary Andersen, a scientist that aided in developing the phylochip, via e-mail last week.
You mention that the waters that are receding from New Orleans after Hurricane Katrina are populated with "bacteria, viruses and other disease-causing microbes." Are there any particular water-borne pathogens that, in your opinion, pose a higher risk to the population over others?
Some organisms such as pathogenic [Escherichia] coli and cholera (Vibrio cholera) are at the top of the list, but the fundemental point with an event like Katrina is what are the unexpected organisms that show up? The main strength of our microarray is a comprehensive identification of the bacteria present in a sample. If the phylochip indentifies a pathogen or close relative of a pathogen, the decision can be made to have a more extensive monitoring for that organism over a larger number of samples.
Could you describe the scenarios that led to the creation of the phlyochip? Whose idea was it and what what set its creation in motion?
This work started when I was a post-doc in Ken Wilson's lab at Duke University Medical Center in 1996. The idea was to develop a method to identify an organism quickly with no prior information. For example, you were given a sample that was said to contain [Bacillus] anthracis (anthrax), but tests showed that this pathogen wasn't present. There was interest to know what was in the sample. Using a 16S ribosomal gene we could start to identify what was there. As technology [improved] to put more DNA probes on smaller surfaces, it was possible to make more sophisticated versions of a 16S ribosomal gene-based identification system possible. That is how over the years I have increased the resolution of organisms that can be detected.
How does the technology work?
Microorganisms in the environment are typically in complex mixtures of hundreds or more individual species. Although DNA-based tests are desirable for identifying particular classes of organisms such as pathogens within these complex mixtures, they are limited in the ability to detect specific DNA sequences of targeted organisms. If one or several pathogens are the only ones of concern, then standard detection technologies are sufficient for monitoring for the targeted pathogens.
If, on the other hand, you do not know what pathogens may be present and desire to know if there is any potential pathogen in your sample (air, water, soil), the technology that we have developed becomes important. With the phylochip that we have developed we are able to simultaneously detect any currently identified organism, including pathogens, on the basis of sequence differences in a specific gene that is possessed by all bacteria, the 16S ribosomal gene. This means that you don't have to go to your sample and guess what organisms are most likely to be there. Instead, you obtain a comprehensive list of all bacteria that are present in the sample.
We use the 16S gene because every bacterium has this gene, it is essential for protein production and therefore essential for life. It is part of a larger protein assembly machinery present in each bacteria and is less likely to move over to another organism by natural recombination in nature. Because of its suitability in bacterial identification, there are over 200,000 16S ribosomal gene sequences in public databases. Variation of 1-3 percent of this 1,500-base-pair sequence is typically the level that bacterial species differ from each other. We can categorize the 16S sequence into about 9,000 different classes, each having a variation of about 1 percent of the 16S gene sequence.
The LBNL press release mentions that the phlyochip could by used to monitor the area affected by Hurricane Katrina. In fact the word "could" is used a lot. How realistic is it that this technology will be adopted?
Obviously, the most important aspect of making this technology adopted by others is to have confidence in the results that it gives. We have validated this phylochip with an extensive collection of air samples obtained from the Department of Homeland Security for sponsored work in the identification of microbial backgrounds that are present in biosurveillance networks. [Andersen has a $678,547, four-year grant for this work — Ed.] Preliminary work with water and soil samples show that this technology is feasible for these environments as well.
We have compared the results of our array with more expensive and time-consuming methods such as the large-scale sequencing of 16S gene library clones and found that we are able to detect almost everything identified by the other methods plus a large number of additional organisms missed by the other techniques. We have verified that the organisms missed by the other method were indeed present in the samples. One of the main issues that we are dealing with now is how to handle the tremendous amount of data generated by this method. We have individual responses from over 500,000 individual probe tests on each array that is used for organism characterization and we usually use multiple arrays for each experiment. This quickly explodes into an overwhelming amount of data.
Even if you can detect pathogens, where does that leave the authorities? How can they respond to microbial identification?
Appropritate measures for containment or treatment are possible if you know what to look for.
When will the chip be ready for use?
We are currently using the chip. There is nothing inherently difficult about using it. The hope is for other labs to use the chip within the next six months. We are already doing many analyses in many types of environments for collegues. The only thing that is left to do is to streamline the analysis software. The output files are large, and if you want to follow specific groups of bacteria over time you have to manually pull this information out of the raw files.