Software Brings Machine Learning to Microscopy

Computational image analysis can be used to detect specific cells in microscopy images automatically.

The system relies on a robust communication channel to read the low-resolution prescan microscope images, interrupt the scan to analyze them and reconfigure the microscope for the desired complex imaging procedure.

Scientists at the European Molecular Biology Laboratory (EMBL; Heidelberg, Germany), developed the platform called the Micropilot, and tested it with three different fluorescence microscopy–based assays, focusing on the biological process of cell division for which low-resolution imaging data was available. The automated system, in a mere four nights of unattended microscope operation, detected 232 cells in two particular stages of cell division and performed a complex imaging experiment on them. Normally, an experienced microscopist would have had to work full-time for at least a month just to find those cells among the many thousands in the sample.

Micropilot can generate enough data to obtain statistically reliable results easily and quickly, allowing scientists to probe the role of hundreds of different proteins in a particular biological process. The scientists at EMBL determined when structures known as endoplasmic reticulum exit sites form, and uncovered the roles of two proteins, in condensing genetic material into tightly wound chromosomes and in forming the spindle, which helps align those chromosomes.

The authors concluded that Micropilot automation reduces bias in cell selection and liberates cell biologists from the tedious work of repetitive manual data generation. It can be adapted to virtually any imaging system that allows automation and online control based on the results of image classification by machine vision. It will be straightforward in the future to use Micropilot to automate additional high-resolution assays beyond multicolor high-resolution confocal three-dimensional stacks, time-lapse movies, fluorescence lifetime imaging microscopy and fluorescence cross-correlation spectroscopy. The study was published online on January 23, 2011, in Nature Methods. The software is available as open source code.

Related Links:
European Molecular Biology Laboratory