Automated image analysis tool to measure microbial growth on solid cultures

Abstract

Microbial growth on culture media is a commonplace technique to estimate the growth rate and virulence of microbes, assess inhibitory effects of compounds and estimate potential damages of plant pathogens in agriculture. Growth area measurement of solid cultures is still commonly performed as a manual process that requires skilled technicians and substantial time, thus warranting an automated system to reduce the workload and increase measurement efficiency. A machine learning approach (Support Vector Machines) was developed to fully automate the area measurement process. We developed a functional model that processes images and returns the microbial area coverage considerably faster than a manual measurement method, with minimal user input and highly comparable results (R2 = 0.88, kappa = 0.88) applicable over large datasets.