Statistical analyses: possible reasons for unreliability of source tracking efforts

Area of Science:

• Microbiology
• Environmental Science
• Genetics

Background:

• Indicator organisms like Escherichia coli (E. coli) are crucial for assessing water's microbiological quality.
• E. coli, a thermotolerant coliform, is typically found in warm-blooded animal feces but can also exist in pristine environmental waters.
• Distinguishing fecal E. coli from environmental E. coli is vital for accurate water quality assessment.

Purpose of the Study:

• To investigate the genetic variability of E. coli isolates from different sources.
• To determine if genetic analysis can reliably differentiate between fecal and environmental E. coli populations.
• To evaluate the effectiveness of pulsed-field gel electrophoresis (PFGE) in source attribution.

Main Methods:

• Comparison of environmental and fecal E. coli isolates against E. coli K12 (ATCC 29425).
• Utilized restriction fragment length polymorphism (RFLP) with pulsed-field gel electrophoresis (PFGE) to analyze genomic DNA patterns.
• Applied cluster analysis and discriminant function analysis for statistical grouping and source attribution.

Main Results:

• PFGE generated distinct band patterns for nearly all analyzed E. coli isolates, indicating high genetic heterogeneity.
• Cluster analysis failed to establish relationships between E. coli isolates and their origin (fecal vs. environmental).
• Discriminant function analysis initially grouped samples by source, but validation revealed this to be a statistical artifact.

Conclusions:

• The significant genetic diversity within E. coli populations complicates reliable source attribution.
• Statistical methods may create apparent groupings that do not reflect true biological separation.
• Accurate differentiation of fecal versus environmental E. coli based on current genetic analysis is challenging.