Their goal is to catalog the thousands of types of bacteria drifting and swirling in the nation's cities, and determine how each bacterium's relative concentration changes week by week. When completed, the census will help researchers differentiate between natural and suspicious fluctuations in airborne pathogens, which can help deter false alarms. It will also help scientists refine tests that identify disease-causing bacteria.

No microbe goes unnoticed as Gary Andersen (left) and Todd DeSantis pore over reams of data that shed light on the nation's population of airborne bacteria.

"If you're developing a way to test for harmful amounts of anthrax, you need to know the background levels of anthrax normally found in the air," says Todd DeSantis of the Earth Sciences Division, who is conducting the research along with project leader Gary Andersen and scientists Jordan Moberg, Sonya Murray, and Ingrid Zubieta.

Although deadly bacteria grab headlines because they can be used as a terrorist weapon, they're also our everyday neighbors. Botulism is caused by the bacterium Clostridium botulinum, which is commonly found in the soil and the intestinal tracts of animals. And anthrax, which is caused by Bacillus anthracis, is an infectious livestock disease. Inevitably, some of these bacteria become aerosolized, meaning they float away in the slightest breeze, and usually in concentrations so small they remain harmless. Their only threat is to confound the search for potentially dangerous concentrations of bacteria, which is why it's necessary to record everything that lurks in the air—the good, the bad, and the esoteric.

The census relies on air samples culled from about 300 air monitors positioned in 30 U.S. cities and surrounding areas. This far-flung network, which monitors the air breathed by about 90 percent of the U.S. population, is one of the nation's first defenses against a terrorist attack. Air samples obtained from the monitors are frequently subjected to a quick, polymerase chain reaction (PCR)-based DNA test designed to detect a handful of dangerous bacteria, such as those that cause anthrax and botulism.

	Bacillus anthracis

Samples that don't contain dangerous pathogens based on the PCR test are then sent to Berkeley Lab where they're exposed to a vastly more comprehensive test capable of detecting 10,000 types of organisms. The test employs a glass chip manufactured by a company called Affymetrix using specifications provided by Berkeley Lab scientists. It boasts a carpet of DNA, with each of the 500,000 carpet strands sporting a section of DNA identical to a section of DNA in one of the 10,000 types of bacteria.

When an air sample arrives at the lab, its bacterial DNA is separated, fluorescently labeled, and washed over the carpet. If DNA from the air sample finds its counterpart on a carpet strand, it attaches itself to the strand, indicating a match. The more matches of DNA specific to a type of bacterium, the more likely that bacterium is in the sample.

"We rely on the statistical significance of many tests coming up positive," says DeSantis. "The trick was choosing the 500,000 pieces of DNA that will test for all 10,000 types of bacteria."

The Berkeley Lab team can churn through 30 air samples in a single week, with each sample representing a week's worth of collected bacteria. They've analyzed hundreds of samples since they started the nationwide background census one year ago, and hope to eventually analyze thousands, enough to cover the entire nation for one year. So far, each sample has yielded hundreds of different organisms from several branches of the taxonomic tree. Many of the bacteria are hardy spores, able to transform from food-gathering organisms in the presence of water to a tight ball of DNA protected by a hard protein coating when water disappears.

"And in many cases we are finding very close relatives of anthrax," says DeSantis. "It emphasizes the importance of designing anthrax tests sensitive enough to differentiate between harmless and harmful types of Bacillus, to avoid false positives."

Escherichia coli   Copyright Dennis Kunkel Microscopy, Inc.

They're also finding E. coli and other organisms that usually reside in animal intestines. Interestingly, these bacteria are more prevalent in urban areas than rural areas, perhaps because there's more sewage and garbage.

The team will soon publish a pilot study that chronicles the ebb and flow of airborne bacteria in San Antonio and Austin over a 20-week span. Feedback from the study will help them determine how best to present the volumes of data they're producing, an important refinement given that a one-week air sample contains about 100 megabytes of data.

They also hope to give any scientist interested in the distribution of airborne bacteria access to their research on the web. For example, a field researcher investigating why birds near a lake are contracting a disease can mine the data to determine which bacteria inhabit upwind regions.

"This is the largest census to date of airborne microbial matter," says DeSantis. "There are other ways to do this, such as cloning and sequencing, but they are more expensive and time-consuming."

Additional information

• More about Berkeley Lab's Center for Environmental Biotechnology