A Survey of Preprocessing Techniques for Flow Cytometry Data in Classification Tasks.

Area of Science:

• Biomedical research
• Clinical diagnostics
• Computational biology

Background:

• Flow cytometry generates complex, multiparametric data essential for disease classification.
• Challenges include spectral overlap, wide dynamic ranges, and batch effects, necessitating preprocessing.
• Current methods often focus on individual cell labeling or patient sample categorization.

Purpose of the Study:

• To provide a comprehensive survey of current preprocessing techniques for flow cytometry data in classification tasks.
• To discuss applications, focusing on signal compensation, batch effect mitigation, imperfect data treatment, and feature selection.
• To emphasize standardizing preprocessing workflows and addressing computational challenges.

Main Methods:

• Survey of existing literature on flow cytometry data preprocessing.
• Categorization of techniques into four key aspects: signal compensation/transformation, batch effect mitigation, imperfect data treatment, and feature selection/class balance.
• Discussion of computational and analytical difficulties.

Main Results:

• Identified key preprocessing areas crucial for flow cytometry data analysis.
• Highlighted the need for standardized workflows to improve model reproducibility.
• Discussed challenges related to data complexity and dataset size.

Conclusions:

• Preprocessing is vital for accurate flow cytometry data classification.
• Standardization and improved methodologies are needed for reproducible results.
• This survey serves as a reference for best practices in flow cytometry data preprocessing.