Cornell University food science, engineering and computer science students have joined forces to develop a web-based software and a database to track and compare genetic footprints, or characteristics, of bacteria.

For scientists who track the spread and sources of virulent bacteria, the students' PathogenTracker software reduces from days and hours to minutes the time spent making tedious strain comparisons.

"Before PathogenTracker, in this laboratory we were using three different databases and two spreadsheets," says Martin Wiedmann, Cornell assistant professor of food science. The food scientists sought the help of the computer science and engineering students, he says, because, "we knew what we wanted but we didn't know how to put the idea together."

The new database is novel in that it allows easy comparisons of strain characteristics and visual images of molecular subtypes (DNA fingerprints). The tool thus allows researchers quickly to assemble strain subtype data from different laboratories in order to analyze outbreaks and epidemics of many different infectious diseases and to assess the biodiversity of bacteria in general.

Before PathogenTracker, scientists used off-the-shelf database programs to track food-borne pathogens. But in 1999 Wiedmann first used his then-primitive database to help limit the death toll from a listeriosis outbreak. Between October 1998 and February 1999, more than 100 people became ill and 21 people died around the country after eating hot dogs contaminated with a rare strain of Listeria monocytogenes -- the deadliest of all food-borne bacteria. Wiedmann's work allowed the Centers for Disease Control (CDC) in Atlanta to determine the cause of the outbreak. Consequently, the contaminated hot dogs were immediately recalled in what was to become one of history's largest food recalls. That early version of PathogenTracker found that seven of the 15 samples had identical genetic fingerprints, meaning that the same strain had caused the illness in seven people. The CDC also had noticed a rise in the number of listeriosis cases, but until Wiedmann's fingerprints, they did not know the strain to look for.

One of the students who developed the new database in Wiedmann's laboratory, Michael Chung '02, began by assembling the pathogen characteristics that the database would need. This included ribotype, DNA sequence and phenotype characteristics. Last fall, Chung began working with a team of computer science students who programmed the software, set up the web server and developed the software's image-recognition capabilities.

By the end of fall 2000, the program was complete. But, says Chung, it could not be transferred onto multiple computers and it was difficult to add new features. Over the past year, Chung, Cornell graduate student Steven Cai, undergraduate student Mike Bohlander '03 and Qi Sun, of the Cornell Computational Biology Service Unit, have greatly improved the program so that it can now be installed on larger web servers and handle a larger load of data and queries.

Currently within the database there are thousands of digital fingerprints, or "isolates," of food-borne pathogens, spoilage organisms and other bacteria such as L. monocytogenes, Pseudomonas, Vibrio parahaemolyticus, Streptococcus and lactic acid bacteria. Thibet Rungrotkitiyot M.E.E. '98, M.E. (industrial) '99, M.E. (computer science) '01, whose experience writing the image-recognition software for the Cornell world champion robotic soccer (RoboCup) team, developed the image recognition software in PathogenTracker that allows the visual comparisons to be made between the genetic fingerprints of the different strains.

Programming for PathogenTracker was developed by Xiaozheng Zhong, '00, M.E. (computer science) '01; David Wang, '00, M.E. (computer science) '01; Joe Cheng-Yu Huang '00, M.E. (computer science) '01; Rungrotkitiyot, Jian-Ning Janet Cheng '01; and Ernie Ho, M.E. (computer science) '01. The development of the library and search engine was completed by Cai, Chung, Wiedmann and staff researcher Roger Jagoda. This project is supported by a special research grant from the U.S. Department of Agriculture and by Dairy Management Inc., by way of Kathryn Boor, Cornell associate professor of food science.