Imaging Platform, Broad Institute of MIT and Harvard, Massachusetts, United States
Abstract: Image-based profiling, also known as morphological profiling, is a rapidly expanding field wherein cells are “profiled” by extracting hundreds to thousands of unbiased, quantitative features from images of cells with genetic or chemical perturbations. It is the least-expensive high-dimensional profiling technique to date, offers single cell resolution, and has been shown to be successful in many varied biological applications including early stage drug discovery.
The Cell Painting assay is the most popular imaged-based profiling assay wherein six small-molecule dyes label eight cellular compartments; after imaging, thousands of conventional measurements describing quantitative traits such as size, shape, intensity, and texture are measured within the nucleus, cytoplasm, and whole cell. Images are also suitable substrates for deep learning, either directly in applications like classification or by use in deep-learning-based feature extraction tools (e.g. DeepProfiler).
We have curated >650 terabytes (TB) of publicly available image data and measurements generated from the Cell Painting assay and close derivatives in the cellpainting-gallery. It is hosted by Amazon Web Services Registry of Open Data. To make the data as FAIR and useable as possible, we have implemented strict data and metadata guidelines, provided comprehensive download instructions, reprocessed old datasets, converted datasets to next-generation-file-formats, and worked with external organizations to make the cellpainting-gallery browseable with their infrastructure.
In this presentation we will describe the cellpainting-gallery, the steps we have taken to make the data FAIR, and how to access the data in the gallery. We will also present several vignettes of biological discovery made possible by datasets hosted in the gallery.