« Previous article
Aflatoxin M1 in Dair...

4th July 2017  Content supplied by: MOCON Inc.

Greenlight® Rapid Microbial Detection System Delivers Accurate CFU Estimate

A laboratory study was conducted to evaluate the key performance parameters of the GreenLight® rapid microbial detection system. The study sought to determine the system’s precision, accuracy, linearity, limit of detection, and limit of quantification. This data could then be used to predict performance in food matrices and allow users to determine the best fit to their application. All tests were performed in an independent ISO17025-certified test laboratory and compared to FDA BAM equivalent methods for Aerobic Plate Count (APC).

The study employed the GreenLight Model 930-15 microbial screening system. GreenLight technology uses a novel polymer sensor to detect changes in the level of oxygen in a sample caused by the growth of aerobic microbes. The oxygen sensor is integrated to the bottom of a consumable sample vial. As microbes grow and respire under incubation, the oxygen in the sample diluent is depleted. This is read as a change in fluorescence from the sensor. Because the measurement is made due to changes in elemental oxygen, there are no limitations due to color, opacity, or turbidity.

Materials and Methods: Precision, accuracy, and linearity testing were done with 4 different bacterial cultures to create a cocktail of E. coli ATCC 25922, L. innocua ATCC 33090, P. aeruginosa ATCC 10145, and S. aureus ATCC 25923. This cocktail started at 7.98 log CFU/mL and was serially diluted in dilution buffer. Each serial dilution had 1mL put into 9 mL of Tryptic Soy Broth in replicates of 5. Plate counts were performed to determine the actual inoculum level. All samples were run and incubated at 32.5C.

Limit of detection and quantification testing were done in the same manner, except only P. aeruginosa ATCC 10145 was used instead of the cocktail. This inoculum started at 7.94 log CFU/mL.

Study: The diluent was pipetted into a 15mL GreenLight sample vial. Due to the dynamic range of the oxygen sensor, no serial dilutions for readability were needed.

The conversion of time to result on GreenLight is a linear correlation equation in the form t= -m(TVC log CFU/g) + T, where t=time to result, m=slope of the linear fit and T=time to intercept at log 0 CFU/g.


  • The Precision Comparison had a Relative Standard Deviation of 1.98% and a Coefficient of Variation of 0.02 at 96 CFU/mL
  • The Accuracy Comparison had an average of 96.89% from 1 to 7 log (CFU)
  • The Linearity Comparison had an R2 value of 0.9533
  • The Limit of Detection was found to be -1.06 log CFU/mL
  • The Limit of Quantification was found to be 0.9 log CFU/mL

The precision comparison shows a relative standard deviation (RSD) of 2.98% for 960 CFU/mL, 1.98% for 96 CFU/mL, and 13.10% for 9.6 CFU/mL. The coefficient of variation for each is 0.03 for 960 CFU/mL, 0.02 for 96 CFU/mL, and 0.13 for 9.6 CFU/mL.

The accuracy of the individual GreenLight log (CFU)/mL values were divided by the plate count log (CFU)/mL values to get an accuracy percentage. The average total for this set from 1 to 7 log (CFU)/mL was 96.89%.

The linearity comparison was completed using a linear regression curve based on standard method plate counting in log (CFU)/mL on the X axis and GreenLight Time to Result (Hours) on the Y axis. The correlation coefficient (R2) was 0.9533 over the range of 1 to 7 log (CFU)/mL.

GreenLight time to result and Plate Count log (CFU)/mL. Graph 1: The linear regression curve (y = -1.3296x + 9.228) of GreenLight time to result and Plate Count log (CFU)/mL.

All 5 replicates reacted at (100%) for the 2.94, 1.94, and 0.94 log (CFU)/mL dilutions reacted. The -0.06 log (CFU)/mL showed 4 out of 5 (80%) reacting and the –1.06 log (CFU)/mL showed 1 out of 5 (20%) reacting.

The testing shows that the instrument gave an accurate CFU estimate 100% of the time all the way down to the 0.9 log (CFU)/mL inoculation level.


  • This test showed a precision with low variation along with excellent accuracy and linearity when compared to standard methods
  • The limit of detection and quantification are better than most standard counting methods

For more information on applications visit www.mocon.com



Date Published: 4th July 2017

Source article link: View

View full company details

Related news