Correlation with plate counts

BactoBox® shows strong agreement with plate count measurements for most Active cultures, i.e. when the culture is in exponential, deceleration, and stationary Growth phases.

Correlation across a growth curve

The below figure compares plate counts (CFU/mL, yellow) with BactoBox® (lavender). Plate counts and BactoBox® curves are practically superimposed, and R² ≈ 1 for the linear regression on the log-log transformed data. This means that the methods are strongly correlated across exponential, deceleration, and early-stationary stage.

K. aerogenes growth curve investigated with BactoBox® (lavender) and plate counts (yellow). Triplicate repeats were done for each data point. More head-to-head growth curves are available in Journal of Microbiological Methods.

Correlation for dilutions of the same sample

The above plot demonstrates the correlation across a growth curve. Correlation between BactoBox® and CFU is also strong across different dilutions of the same sample as demonstrated on the plot below where the same sample was diluted and measured head-to-head with BactoBox® and plate counts.

The slope of the regression line is 0.9947. This is practically a perfect slope of one meaning that for every S. epidermidis detected on BactoBox® you will also get one colony by plate counts. The R² ≈ 1 signifies the strong linear relationship between the two methods. In most cases, BactoBox® demonstrates greater precision than the CFU method. For example the most concentrated sample has a coefficient of variation (CV) of 3 % for BactoBox, while the CV is 35 % for the plate count method.

Head-to-head comparison of BactoBox® and plate count method for dilutions of the same Staphylococcus epidermidis sample. X-axis represents plate count arithmetic mean and y-axis represents BactoBox® arithmetic mean. All measurements were performed in technical triplicates. The table displays the arithmetic mean and the coefficient of variation (CV) for each data point. CV ≤ 5 % are highlighted in green.