Skip to Content

Food-Borne Bug Hunter

Researchers at Cornell University have developed an Internet database that identifies bacterial strains in minutes.
February 1, 2002

By many measures, the U.S. food supply is the safest in the world. Nevertheless, the U.S. Centers for Disease Control and Prevention (CDC) estimates that each year about 70 million Americans-nearly one quarter of the total U.S. population-contract some sort of food-borne illness. While the vast majority of these cases are minor, approximately 300,000 require hospitalization, and 5,000 end in death annually.

Now students and faculty at Cornell University have assembled a Web-based library containing the genetic fingerprints of food-borne pathogens-aptly called PathogenTracker-which can help health officials rein in food-borne outbreaks. Because each bacterial strain has a unique genetic fingerprint, private labs can compare their own bacterial samples to those in the database, then quickly determine whether, for example, multiple incidents of food poisoning are coincidental or share a common source. If strains from different patients match, public-health officials can quickly track down the contaminated food source and remove it from the market, halting a wider outbreak.

The database currently contains information on thousands of strains of six different bacterial species, and that number will soon grow. The strength of PathogenTracker, claims the group, is that labs from all over the world can submit data on newly isolated strains of bacteria, and there is virtually no limit to how much information the database can contain. “This database will never be done. It will continually be updated and modified,” says Martin Wiedmann, assistant professor of food science at Cornell and the group’s leader.

Identifying the Usual (Bacterial) Suspects

Traditionally, public health workers have identified outbreaks of food poisoning by focusing on specific events, such as a church picnic where one bad batch of chicken salad infects fifty people in a single afternoon. But it is far more difficult to detect lower-level outbreaks that infect a few people over a wide area. For example, in 1993 it took health officials about a week to recognize dozens of reported illnesses in the Pacific Northwest as an E. coli outbreak, and another week to identify the source: hamburger meat from the Jack-in-the-Box fast-food chain. Meanwhile, 720 people fell ill, and four died.

This prompted the CDC in 1996 to build its own centralized database of food-borne pathogens, called PulseNet. Labs submit the genetic fingerprints of strains that they’ve isolated to the CDC. As each participating lab submits a sample, the CDC records it and is thus able to maintain nation-wide surveillance. The agency also sends back to the lab the identity of the strain that was submitted.

Despite some initial doubt among researchers about the project’s feasibility, PulseNet has proved highly effective in the early detection of outbreaks since its 1998 launch. Currently 64 labs in the U.S. and Canada participate, and the database contains nearly 20,000 unique fingerprints from a total of four bacterial species.

“We are preventing people from getting sick and consequently we are saving lives,” says Bala Swaminathan, chief of food-borne diseases laboratories at the CDC and lead developer of PulseNet.

But PulseNet is not without limitations. First, the system accommodates only one pathogen fingerprinting method: pulsed-field gel electrophoresis. In this process, a pathogen’s genome is divided into eight to 20 pieces and then placed in a gel. Electric pulses cause the DNA to unravel and separate into a unique pattern based on the size of the fragments.

While this is the most broadly applicable method, it isn’t always the most sensitive, and labs use about a dozen other methods to test for different organisms. In addition, only those labs that pass the CDC’s rigorous screening process can contribute to and query PulseNet. And even when participating labs send in data on their samples, the CDC takes a minimum of 24 hours to respond-a vast improvement over previous methods, but still not real-time.

Pathogen Tracking Goes Public

The Cornell group’s system seeks to fill the gaps left by PulseNet. Researchers can query the database with pathogen fingerprints created by a variety of methods. Anyone with a Web browser can submit new bacterial fingerprints (which PathogenTracker’s managers review before posting to maintain quality) or search the library, and the system returns results in a matter of minutes. Researchers can even query the database anonymously.

But the key ingredients, according to Wiedmann, are versatility and time. When hospitals and labs analyze a bacterial sample from a patient, they can use the method they feel is most appropriate to generate a genetic fingerprint and then compare it to the PathogenTracker database. Because they get results immediately, the contaminated food can be tracked down that much faster.

Currently the library contains data on thousands of bacterial samples. For example, PathogenTracker contains 3,660 individual samples of Listeria monocytogenes, a deadly pathogen often found in ready-to-eat foods like hot dogs, deli meats and soft cheeses. In addition to genetic fingerprints for each sample, the system also contains basic statistical information, such as where the bacteria were first isolated-without identifying specific establishments or patients.

The real hope behind PathogenTracker is that it will draw together research labs from all over the world, says Wiedmann, truly becoming a public project whose “limitless capacity” for amassing isolates can be put to the test.

Joining Forces

Wiedmann, however, has no intention of replacing PulseNet. The two aren’t competitors, nor should they be. While the CDC’s certification system is limiting for many labs, it also guarantees a certain level of quality to its database. PulseNet also allows the CDC to maintain nationwide surveillance. As labs across the country submit their data, the CDC can get some indication of whether a seemingly isolated case in, say, Montana, is related to a case in Florida-a level of observation that the PathogenTracker team cannot provide if labs choose to query the database anonymously.

“No single data system is going to give you all that’s available,” says Walter Hill, director of the bioscience division at the U.S. Department of Agriculture’s Food Safety and Inspection Service, which has partially funded Wiedmann’s research. To help fill remaining gaps, the CDC and Wiedmann often collaborate. In 1998, for example, an early version of PathogenTracker helped the CDC identify the source of a Listeria outbreak that killed 21 people across the country.

Wiedmann plans to make PathogenTracker publicly available in the next month or two. And if the project takes off, it should be that much easier to stop a food poisoning outbreak right in its tracks.

Keep Reading

Most Popular

Geoffrey Hinton tells us why he’s now scared of the tech he helped build

“I have suddenly switched my views on whether these things are going to be more intelligent than us.”

Meet the people who use Notion to plan their whole lives

The workplace tool’s appeal extends far beyond organizing work projects. Many users find it’s just as useful for managing their free time.

Learning to code isn’t enough

Historically, learn-to-code efforts have provided opportunities for the few, but new efforts are aiming to be inclusive.

Deep learning pioneer Geoffrey Hinton has quit Google

Hinton will be speaking at EmTech Digital on Wednesday.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at with a list of newsletters you’d like to receive.