Stitching and registering highly multiplexed whole slide images of tissues and tumors using ASHLAR software

Abstract

Widespread use of highly multiplexed microscopy to study normal and diseased tissues at a single-cell level is complicated by underdevelopment of the necessary software. This is particularly true of high resolution whole-slide imaging (WSI), which involves gigapixel datasets of specimens as large as 5 cm2. WSI is necessary for accurate spatial analysis and a diagnostic necessity. High resolution WSI requires collection of successive image tiles; multiplexing commonly involves successive data acquisition cycles, each with a subset of dyes, antibodies or oligonucleotides. We describe a new Python tool, ASHLAR (Alignment by Simultaneous Harmonization of Layer/Adjacency Registration), that coordinates stitching and registration and scales to 103 or more image tiles over many imaging cycles to generate accurate, high-plex image mosaics, the key type of data for downstream visualization and computational analysis. ASHLAR is more robust and accurate than existing methods and compatible with any scanner or microscope conforming to Open Microscopy Environment standards.