Statistical interpretation of the Amerithrax "morph" assay results

Area of Science:

• Forensic Science
• Statistical Modeling
• Molecular Biology

Background:

• PCR-based "morph assays" were critical in the Amerithrax investigation to link anthrax letters to the RMR-1029 flask at USAMRIID.
• Quantitative data from these assays deviated from expected Poisson statistics, showing "Taylor's Law" behavior with variance exceeding the mean.

Purpose of the Study:

• To develop a statistical model explaining the observed assay behavior and its implications for interpreting results.
• To assess the impact of assay limitations on the conclusions drawn in the Amerithrax investigation.
• To establish a rigorous statistical framework for interpreting "morph" assay data.

Main Methods:

• Analysis of quantitative PCR data using "Taylor's Law" to model variance-mean relationships.
• Development of a statistical model to explain negative and "inconclusive" assay findings.
• Application of a Bayesian framework to relate assay results to the probability of sample contamination.

Main Results:

• The "Taylor's Law" behavior explains a high number of negative and "inconclusive" results, suggesting "morph" mutants can be present but undetected.
• A repository sample could contain "morph" mutants above detection limits yet yield negative or inconclusive results.
• The Bayesian analysis indicates that the false negative rate does not significantly undermine the exclusion of non-RMR-1029 sources for unambiguous assay results.

Conclusions:

• The statistical model provides a plausible explanation for observed "morph" assay anomalies in the Amerithrax investigation.
• The findings resolve issues raised in recent reviews regarding assay interpretation.
• Developing a robust statistical framework is crucial for accurate interpretation of "morph" assay data in forensic investigations.