Piccinini, F. et al. Cell Syst. 4, 651–655.e5 (2017).

High-throughput and high-content microscopy data are invaluable for biological research, yet they are generated at a scale that can preclude fully manual interrogation. A number of algorithmic tools, such as machine-learning approaches, have been developed for semiautomated or fully automated assessment of large image data sets. Piccinini et al. describe a second version of their Advanced Cell Classifier software package for phenotypic analysis of large-scale, cell-based microscopy experiments. The software is meant to provide a user-friendly package to facilitate training machine-learning algorithms and mine data to classify cells, discover new and/or rare phenotypes, and improve the accuracy of the analysis process. The software package was validated on synthetic and experimental data, and it should prove useful for researchers working in high-content imaging.