CYTO 2026 Scientific Tutorial: Crimes Against Data Analysis

Crimes Against Data Analysis

Presenters:
Sarah Bonte, PhD, Postdoctoral Researcher, VIB-Ghent University
Geoffrey Kraker, BSc, Senior Application Specialist, Dotmatics
Givanna Putri, PhD, Postdoctoral Researcher, Walter Eliza Hall Institute of Medical Research
Nicolas Loof, MSc, Informatics Solution Leader, BD/FlowJo

Abstract:

High parameter cytometry technologies enable simultaneous characterization of many cellular markers, offering unprecendented insights into complex biological systems. As panel sizes and complexities continue to expand with advances in instruments and reagents, our capacity to analyze data has not kept pace, making high-dimensional data increasingly challenging to interpret. While data analysis tools are becoming more sophisticated and accessible, whether through programming platforms or commercial software plugins, longstanding issues with data quality remain a significant barrier. These are further compounded by artifacts introduced during data processing, which can lead to erroneous interpretations. Yet, clear guidance on how to recognize, diagnose, and mitigate them is still lacking.

This tutorial will provide strategic and tactical approaches for identifying and troubleshooting common data quality issues and artifacts in cytometry data, showing both their causes and effects. Most commonly used algorithms for high-dimensional data analysis will be covered, along with guidelines for parameter settings for each.

Learning objectives:

Having an idea of the output of computational algorithms when there are no/minimal problems with data quality and no artifacts introduced during data analysis (""What it should look like"")
Recognizing artifacts introduced by data analysis and/or data quality issues (""What it looks like if you don't do it right"")
Knowing what to do to prevent these artifacts from occurring
Understanding parameter choices in commonly used computational algorithms for high-dimensional data analysis
Guidelines on how to pick the most optimal parameters for your data, and diagnose when you have picked the wrong ones

Keywords: CytoBytes, Clustering, High Parameter

CMLE Credit: 1.5