Related Experiment Videos
Comparison of statistical methods for identification of Streptococcus thermophilus, Enterococcus faecalis, and
G Moschetti1, G Blaiotta, F Villani
1Dipartimento di Scienza degli Alimenti, Università degli Studi di Napoli Federico II, 80055 Portici, Italy.
Applied and Environmental Microbiology
|April 25, 2001
Summary
A new study shows that Bayesian networks can quickly and affordably identify thermophilic streptococci, like Streptococcus thermophilus, using simplified Randomly Amplified Polymorphic DNA (RAPD) PCR patterns.
Area of Science:
- Microbiology
- Bioinformatics
- Food Science
Background:
- Thermophilic streptococci are crucial in European cheese production.
- A need exists for rapid and reliable identification methods for these bacteria.
Purpose of the Study:
- To develop a fast, inexpensive method for identifying thermophilic streptococci.
- To evaluate the effectiveness of artificial neural networks and statistical techniques for bacterial identification.
Main Methods:
- Generated Randomly Amplified Polymorphic DNA (RAPD) PCR patterns using a single primer (XD9).
- Analyzed patterns using artificial neural networks (Bayesian network, Multilayer Perceptron, Radial Basis Function) and statistical methods (cluster analysis, linear discriminant analysis, classification trees).
Main Results:
- Cluster analysis identified Streptococcus thermophilus but not enterococci.
- A Bayesian network was more effective than other networks and statistical methods for identifying Streptococcus thermophilus, Enterococcus faecium, and Enterococcus faecalis.
- The Bayesian network performed well with simplified RAPD-PCR patterns and was robust to training set size and unknown species.
Conclusions:
- Bayesian networks offer a superior method for identifying thermophilic streptococci compared to traditional statistical techniques.
- Simplified RAPD-PCR patterns analyzed by Bayesian networks provide a robust and efficient identification tool for food microbiology.