Toward reproducible, scalable, and robust data analysis across multiplex tissue imaging platforms

Abstract

Summary: The emergence of megascale single-cell multiplex tissue imaging (MTI) datasets necessitates reproducible, scalable, and robust tools for cell phenotyping and spatial analysis. We developed open-source, graphics processing unit (GPU)-accelerated tools for intensity normalization, phenotyping, and microenvironment characterization. We deploy the toolkit on a human breast cancer (BC) tissue microarray stained by cyclic immunofluorescence and present the first cross-validation of breast cancer cell phenotypes derived by using two different MTI platforms. Finally, we demonstrate an integrative phenotypic and spatial analysis revealing BC subtype-specific features. Motivation: Multiplex tissue imaging (MTI) is transforming biology by linking biomarker expression with spatial context. As the number of MTI platforms and spatial tissue atlases increases, so too must the capacity of our methods to ingest and derive insight from these data. Many of the methods used in MTI analysis are crossovers from other domains like flow cytometry and single-cell genomics, do not scale well, and might be statistically unfit for atlas-level MTI data, constituting a crucial bottleneck in the search for the next generation of clinically relevant biomarkers. Here, we address this bottleneck with a computationally efficient and platform-agnostic MTI data analysis workflow and its proof-of-concept application to human breast cancer (BC) tissues.

Keywords

• multiplex tissue imaging
• GPU data science
• breast cancer
• systems biology
• image analytics