Calculating the limit of detection for a dilution series

Area of Science:

• Microbiology
• Statistical Modeling
• Biofilm Research

Background:

• Microbial quantification often involves serial dilutions to estimate cell numbers.
• Existing methods for determining the limit of detection (LOD) in dilution series can be simplistic.
• A statistically robust definition of LOD is crucial for accurate microbial enumeration.

Purpose of the Study:

• To define and statistically calculate the limit of detection (LOD) for microbial counts from dilution experiments.
• To extend existing chemical detection limit methodologies to microbiological applications.
• To provide a practical and accessible method for determining microbial LOD.

Main Methods:

• Utilized the negative binomial distribution to model microbial counts, overcoming limitations of the Poisson distribution.
• Developed a statistical definition for LOD based on high detection probability (e.g., 0.95).
• Illustrated the method using data from Pseudomonas aeruginosa biofilms.

Main Results:

• The LOD is determined by factors including statistical power, overdispersion, lowest countable dilution, sample volume, and number of replicates.
• The negative binomial model provides a more accurate estimation than Poisson-based approaches.
• Demonstrated practical application in a real-world biofilm study.

Conclusions:

• A statistically sound and accessible method for calculating the LOD in microbial dilution series has been established.
• This approach enhances confidence in quantifying microbial populations.
• The methodology is applicable to any zero-count observation process across scientific disciplines.