optimalFlow

optimalFlow is an R package that addresses the challenges of supervised gating in flow cytometry data, characterized by significant biological and technical variability. The software introduces optimalFlowTemplates, a method based on similarity distance and Wasserstein barycenters, to cluster cytometries and generate prototype cytometries for different groups. This approach enhances supervised learning by restricting it to the new groups defined by optimalFlowTemplates, demonstrating improved performance compared to techniques applied to the entire dataset. The package also includes optimalFlowClassification, utilizing a database of gated cytometries and optimalFlowTemplates to assign cell types to new cytometries.