Cyto-Feature Engineering

"Cyto-Feature Engineering" is a flow cytometry analysis pipeline to address the challenges of analyzing high-dimensional data generated by modern flow cytometers, which can analyze up to 50 parameters per cell across millions of cells per sample. Traditional methods of flow cytometry data analysis often involve subjective and time-consuming processes. This R-based pipeline introduces a systematic and efficient approach to identifying various cell populations, incorporating feature engineering and providing immunological context through intuitive plots.

Key Features and Functionalities:

- Automated Threshold Development: Utilizes Fluorescence Minus One (FMO) controls or identifies distinct population differences to automate the development of thresholds for marker expression, transitioning continuous data into binary data to reflect the positive/negative biological dichotomy.

- Data Filtering and Refinement: Incorporates a filtering step that refines the data from all identified cell phenotypes to focus on populations of interest, streamlining the analysis process.

- Modular Design for Customization: The pipeline is modular, allowing researchers to customize statistical testing, adapt initial gating steps, and incorporate additional datasets as needed.

- Statistical Correlation and Partitioning: Enables data partitioning by immune lineages and facilitates statistical correlation with other experimental measurements, enhancing the interpretability of flow cytometry data in research contexts.

- Validation and Accuracy: Validation using manually gated datasets, including murine splenocytes and human whole blood, confirmed the pipeline's accuracy in identifying cell populations, including rare subsets.

- Wide Applicability: Designed to be applicable across all disciplines utilizing flow cytometry, the pipeline is compatible regardless of the cytometer or panel design, making it a versatile tool for a wide range of flow cytometry applications.