Quality Control Measures and Statistical Strategies to Address the Challenges of High-Content Phenotypic Data

High-content image-based cytological profiling is a powerful strategy for studying the effects of chemical and genetic perturbations on the cell. Cytological profiling assays illuminate multiple cellular compartments within each cell by multiplex fluorescent staining, followed by automated microscopy and image analysis. In this chapter, we show how to utilize data derived from images of fluorescently labeled cells and organelles while simultaneously addressing common challenges of this data type. We discuss different modes of interpreting raw cellular features and describe statistical methods for using said features to quantitatively evaluate overall assay quality and reproducibility. Data standardization is described as a two-tiered task, and the more recent EMD metric is implemented as a quantitative measure of phenotypic change. We illustrate each technique using data from an osteosarcoma (U-2 OS) high-content screening assay and describe all tools required to reproduce this work.