Image content-based retrieval and automated interpretation of fluorescence microscope images via the protein subcellular location image database

We describe the Protein Subcellular Location Image Database (PSLID), which collects and structures 2-D through 5-D fluorescence microscope images, annotations, and derived features in a relational schema. It is designed so that interpretations as well as annotations can be queried. The annotations in PSLID, composed of 28 linked tables with publicly available descriptions, provide a thorough description of sample preparation and fluorescence microscope imaging. Image interpretation is achieved using Subcellular Location Features that have been shown to be capable of recognizing all major subcellular structures and of resolving patterns that cannot be distinguished by eye. Results of queries can be ranked by image typicality and statistical tests can be performed on sets of images (e.g., to determine whether a drug alters the distribution of a tagged protein). We anticipate that PSLID will serve as a common repository for microscopy images documenting the subcellular location of proteins.