CMStatistics 2021: Start Registration
View Submission - CMStatistics
B0798
Title: Computational tools to quantify and correct slide-to-slide variation in multiplexed immunofluorescence images Authors:  Coleman Harris - Vanderbilt University Medical Center (United States) [presenting]
Eliot McKinley - Vanderbilt University Medical Center (United States)
Joseph Roland - Vanderbilt University Medical Center (United States)
Qi Liu - Vanderbilt University Medical Center (United States)
Martha Shrubsole - Vanderbilt University Medical Center (United States)
Ken Lau - Vanderbilt University Medical Center (United States)
Robert Coffey - Vanderbilt University Medical Center (United States)
Julia Wrobel - Colorado School of Public Health (United States)
Simon Vandekar - Vanderbilt University (United States)
Abstract: The multiplexed imaging domain is a nascent single-cell analysis field with a complex data structure susceptible to technical variability that disrupts inference. These in situ methods are valuable in understanding cell-cell interactions, but few computational tools exist to quantify or correct for technical variation in multiplexed imaging data. We implement and compare normalization algorithms in multiplexed imaging data, and present and evaluate the methods with a new R package, MxNorm. Our methods adapt the ComBat and functional data registration methods to remove slide effects in this domain, and we include a robust evaluation framework to compare the proposed approaches. The methods illustrate clear slide-to-slide variation in the raw, unadjusted data, demonstrating that many of the proposed normalization methods reduce this variation while preserving and improving the biological signal. The R package introduced here provides a clear framework to normalize and explore multiplexed imaging data, with methods to compare normalization algorithms and visualization tools to identify technical variation. Further, this framework can be integrated with any proposed normalization method or thresholding algorithm, providing open-source software to robustly improve data quality and evaluation criteria in the multiplexed domain.