cyCONDOR: A Comprehensive Workflow for High-Dimensional Cytometry Data Analysis
cyCONDOR is a versatile and powerful workflow designed for data pre-processing and analysis in high-dimensional cytometry (HDC). It offers a suite of microservices that support a wide range of data input formats, including Flow Cytometry Standard (FCS) and Comma-Separated Values (CSV) files, making it compatible with data exported from popular flow cytometry software like FlowJo. The platform also facilitates the import of FlowJo workspaces, enabling users to leverage existing gating strategies for seamless integration of conventional and cluster-based cell annotation.
Data Pre-Processing and Visualization
The workflow begins with data importation, followed by essential pre-processing steps to ensure compatibility with downstream analyses. Users can choose between importing all recorded events or applying broad gating to reduce dataset size. Basic gating is recommended to exclude debris and doublets, optimizing computational efficiency. cyCONDOR provides comprehensive tools for exploratory data analysis, including principal component analysis (PCA) and dimensionality reduction techniques such as Uniform Manifold Approximation and Projection (UMAP) and t-distributed Stochastic Neighbor Embedding (tSNE). These tools help visualize data structure and identify patterns, enabling users to explore biological differences and relationships between samples.
Clustering and Annotation
cyCONDOR integrates two clustering algorithms, Phenograph and FlowSOM, to enable both knowledge-based and data-driven identification of cell lineages and novel cell states. Phenograph provides fine-grained clustering, while FlowSOM offers faster, hierarchical clustering. The platform also includes automated heatmap visualization of marker expression, aiding in the biological annotation of clusters. Users can manually label clusters based on prior knowledge, setting the stage for further downstream analysis.
Batch Correction and Scalability
HDC datasets often suffer from technical variability across batches, instruments, or time points. cyCONDOR addresses this challenge by implementing Harmony, a batch correction method originally developed for single-cell RNA sequencing. Harmony normalizes technical variance, allowing for effective integration of datasets across different sources. This feature is particularly valuable for longitudinal studies or multi-center trials where maintaining consistency across datasets is critical.
Trajectory Inference and Developmental Analysis
cyCONDOR extends its functionality to trajectory inference, enabling the exploration of continuous developmental processes within cellular populations. Using the slingshot algorithm, the platform can predict pseudotime and infer trajectories, providing insights into cellular differentiation and lineage progression. This capability is demonstrated through the analysis of hematopoietic stem cell differentiation pathways, showcasing the utility of pseudotime analysis in HDC data.
Differential Analysis and Statistical Testing
The platform simplifies comparisons between experimental groups through built-in functions for statistical testing and visualization. Users can perform differential abundance and differential expression analyses using tools like diffcyt, while visualizations such as stacked bar plots and heatmaps provide intuitive insights into biological differences between groups. These features are essential for identifying significant changes in cell frequencies and marker expression across conditions.
Machine Learning and Clinical Classification
cyCONDOR incorporates machine learning capabilities for clinical classification, leveraging tools like CytoDx to predict clinical outcomes. The platform can train classifiers at both the single-cell and sample levels, enabling accurate diagnosis and prognosis. By integrating these features, cyCONDOR supports the translation of HDC data into clinically relevant insights, with applications ranging from hematological malignancies to infectious diseases.
Conclusion
cyCONDOR is a robust and scalable ecosystem for HDC data analysis, offering a comprehensive suite of tools for pre-processing, clustering, batch correction, trajectory inference, and machine learning. Its flexibility and user-friendly design make it accessible to researchers and clinicians alike, facilitating the extraction of valuable biological and clinical insights from complex cytometry datasets.
