High throughput microscopy: from images to data

For academic scientists, high throughput microscopy enables a paradigm shift from labor intensive qualitative microscopy to fast image-based quantitative biology. For pharmaceutical scientists, it enables replacing whole-well averaged readouts of simple cellular biomarkers with multiplexed cellular systems biological screens that include cell subpopulation and morphological characteristics. For research pathologists, the future looks bright for automating the process of reading and scoring tissue microarrays, which consist of a few hundred to a few thousand tissue sections on each slide. For image-based (or high content) screening, various algorithms have been developed to perform the automated cytometry needed to assess compound hits and/or dose responses. For many other drug screening and academic biological assays and for tissue microarrays, substantial challenges remain to be solved by more advanced image analysis techniques. For all of these applications, raw image data can comprise tens to hundreds of gigabytes (GB) per experiment and new frameworks enabling rapid algorithm development and validation are needed