Machine learning for enumeration of cell colony forming units

Area of Science:

• Microbiology
• Bioinformatics
• Computational Biology

Background:

• Bacterial cell enumeration is crucial for biological research but traditionally time-consuming.
• Manual colony counting is labor-intensive and prone to human error.

Purpose of the Study:

• To develop and validate a machine learning-based automated method for counting bacterial colony forming units (CFUs).
• To improve the efficiency and accuracy of bacterial enumeration in biological assays.

Main Methods:

• Development of CFUCounter, a program utilizing unsupervised machine learning, iterative adaptive thresholding, and watershed segmentation.
• Processing of digital images to segment and count bacterial colonies.
• Comparison of CFUCounter performance against manual counting and point-and-click methods.

Main Results:

• CFUCounter demonstrated high accuracy and robustness in enumerating CFUs, with performance metrics (slope 0.996, r=0.999) indistinguishable from gold standards.
• The algorithm successfully segmented bacterial colonies from digital images.
• CFUCounter supports color-based classification and counting of heterologous colonies.

Conclusions:

• CFUCounter offers an accurate, efficient, and automated solution for bacterial colony counting.
• This open-source tool provides a valuable addition to existing methods for microbial enumeration.
• The CFUCounter application streamlines a critical but laborious step in biological research.