Digital Image Processing for High Throughput Microbiological Assays: Implementing Customizable Software to Streamline Data Processing

Typical in vitro microbiology assays that produce high volumes of agar plates for colony counting can be overwhelming, time-consuming, and inaccurate when plates are counted manually. To cut down on overall plate counts, iFyber uses a high throughput method that employs a streamlined multichannel-aligned spot plating technique along with traditional spread plating. Even with this streamlined approach, typical projects produce hundreds of plates per day that require CFU counting—where, often, plates contain CFUs at a level of 1000s per plate that would be considered “too many to count” (TMTC) for practical purposes.

With advances in image processing software, iFyber has boosted its project throughput by implementing an automated, accurate solution via MIPAR Image Analysis. MIPAR’s user-friendly interface allows for greater flexibility, throughput, and customizability when compared to open-source competitor platforms such as ImageJ and OpenCV that would be impractical for large plate volumes. iFyber has successfully applied an automated spread-plate-counting recipe and a semi-automated spot-plate-counting recipe in MIPAR, increasing assay throughput and accuracy and, importantly, implementing better traceability for our research results.

Briefly, batches of agar plates are imaged on a flatbed scanner, and the saved image is loaded into the MIPAR platform, which detects, segments, and measures the amount of CFUs on each plate, and may be applied in the MIPAR Batch Processor to process multiple images at once. Custom MIPAR recipes are utilized for both spot plates and spread plates to count colonies of bacteria (many gram negative and positive species) and yeast. By removing the majority of the human element in colony counting, iFyber has implemented an innovative technological solution that reduces counting errors and, importantly, the mental load of manually counting colonies from hundreds of plates per day. Further, our approach reduces the time to answer and provides a digital fingerprint of microbiological data.