Detailed description of the image analysis workflow, including machine learning model training dataset, for the article: "Phosphate starvation decouples cell differentiation from DNA replication control in the dimorphic bacterium Caulobacter crescentus", Hallgren et al. (2023; PLOS Genetics).

More specifically, the procedure identifies individual C. crescentus cells in phase-contrast microscopy pictures, annotates their cell type, as well as their size and basic morphological features. The procedure can for example be used to quantify the proportion of swarmer cells to stalked cells in a population, to measure the size of cells and their stalks, and to determine the fraction of constricted predivisional cells in a population.

The dataset includes: (1) ilastik project files, (2) custom Python scripts and ImageJ macros used for data processing, (3) settings for batch processing of images with the ImageJ plugin ‘MicrobeJ’, and (4) example data that can be used to run the image analysis pipeline from start to finish. The ilastik project files (.ilp) contain the random forest machine learning models used for their analysis, as well as the training dataset used to generate those models.