{"id":395,"date":"2022-12-01T11:31:20","date_gmt":"2022-12-01T10:31:20","guid":{"rendered":"https:\/\/paul-regnier.fr\/?page_id=395"},"modified":"2026-01-29T10:29:33","modified_gmt":"2026-01-29T09:29:33","slug":"tutoriel-picaflow","status":"publish","type":"page","link":"https:\/\/paul-regnier.fr\/en_gb\/tutoriel-picaflow\/","title":{"rendered":"PICAFlow tutorial"},"content":{"rendered":"

PICAFlow: Pipeline for Integrative and Comprehensive Analysis of flow\/mass cytometry data<\/strong><\/p>\n\n\n\n

Warning: this tutorial is only available in English, even if you choose the French language at the bottom of the screen. Thank you for your understanding.<\/strong><\/p>\n\n\n\n

PICAFlow<\/code> is a R package allowing to process cytometry data from raw FCS files to deep and comprehensive analysis of underlying key messages.<\/p>\n\n\n

\n
\"\"\n\t\t\t\n\t\t\t\t\n\t\t\t<\/svg>\n\t\t<\/button>
An AI-generated visual representation of what PICAFlow<\/code> could be in real life. Drawn by Craiyon.<\/strong><\/figcaption><\/figure>\n<\/div>\n\n\n

Table of contents<\/strong><\/h2>\n\n\n\n
    \n
  1. Prerequisites<\/a>\n
      \n
    1. PICAFlow R package installation<\/a><\/li>\n\n\n\n
    2. Troubleshooting<\/a><\/li>\n\n\n\n
    3. Load PICAFlow<\/a><\/li>\n\n\n\n
    4. Update PICAFlow<\/a><\/li>\n\n\n\n
    5. Example dataset<\/a><\/li>\n<\/ol>\n<\/li>\n\n\n\n
    6. Pre-processing<\/a>\n
        \n
      1. Working directory setup<\/a><\/li>\n\n\n\n
      2. Convert FCS files to rds data files<\/a><\/li>\n\n\n\n
      3. Pre-gating (optional)<\/a>\n
          \n
        1. Merge samples<\/a><\/li>\n\n\n\n
        2. Gate cells<\/a><\/li>\n\n\n\n
        3. Reexport individual rds files<\/a><\/li>\n<\/ol>\n<\/li>\n\n\n\n
        4. Subset data<\/a><\/li>\n<\/ol>\n<\/li>\n\n\n\n
        5. Transformation and compensation<\/a>\n
            \n
          1. Visually determine optimal transformation parameters<\/a><\/li>\n\n\n\n
          2. Apply transformation<\/a><\/li>\n\n\n\n
          3. Visually edit compensation matrix (optional)<\/a><\/li>\n\n\n\n
          4. Compensate data (optional)<\/a><\/li>\n\n\n\n
          5. Export data per parameter<\/a><\/li>\n<\/ol>\n<\/li>\n\n\n\n
          6. Normalization<\/a>\n
              \n
            1. Plot transformed signal densities<\/a><\/li>\n\n\n\n
            2. Peaks analysis<\/a><\/li>\n\n\n\n
            3. Compute peaks data for new samples but keep previously generated peaks data (optional)<\/a><\/li>\n\n\n\n
            4. Fine tune the peaks interactively (optional)<\/a><\/li>\n\n\n\n
            5. Synthesize peaks information<\/a><\/li>\n\n\n\n
            6. Normalize data<\/a><\/li>\n\n\n\n
            7. Plot normalized signal densities<\/a><\/li>\n\n\n\n
            8. Merge normalized data<\/a><\/li>\n<\/ol>\n<\/li>\n\n\n\n
            9. Gating<\/a><\/li>\n\n\n\n
            10. Downsample, rescale and split data<\/a>\n
                \n
              1. Downsample and rescale data<\/a><\/li>\n\n\n\n
              2. Split data (optional)<\/a><\/li>\n<\/ol>\n<\/li>\n\n\n\n
              3. UMAP dimensionality reduction<\/a>\n
                  \n
                1. Initial preparation<\/a><\/li>\n\n\n\n
                2. Determination of optimal UMAP hyperparameters<\/a><\/li>\n\n\n\n
                3. Generate UMAP model on downsampled dataset<\/a><\/li>\n\n\n\n
                4. Apply UMAP model to the remaining data<\/a><\/li>\n\n\n\n
                5. Combine and export UMAP data<\/a><\/li>\n<\/ol>\n<\/li>\n\n\n\n
                6. Export FCS files<\/a><\/li>\n\n\n\n
                7. Clustering<\/a>\n
                    \n
                  1. Test parameters normality<\/a><\/li>\n\n\n\n
                  2. Apply a clustering method<\/a>\n
                      \n
                    1. Using hierarchical clustering + k-nearest neighbors approach<\/a>\n
                        \n
                      1. Initial clustering on training dataset<\/a><\/li>\n\n\n\n
                      2. Final clustering on training dataset<\/a><\/li>\n\n\n\n
                      3. Apply clustering model on validation dataset<\/a><\/li>\n\n\n\n
                      4. Determine binary thresholds<\/a><\/li>\n\n\n\n
                      5. Collapse phenotypically close clusters<\/a><\/li>\n<\/ol>\n<\/li>\n\n\n\n
                      6. Using FlowSOM method<\/a><\/li>\n\n\n\n
                      7. Using PhenoGraph method<\/a><\/li>\n<\/ol>\n<\/li>\n\n\n\n
                      8. Visualize clusters on UMAP dimensions<\/a><\/li>\n\n\n\n
                      9. Export clusters-associated statistics and plots<\/a><\/li>\n\n\n\n
                      10. Export clusters-associated heatmaps<\/a><\/li>\n<\/ol>\n<\/li>\n\n\n\n
                      11. Metadata integration and analysis<\/a>\n
                          \n
                        1. Open and merge datasets<\/a><\/li>\n\n\n\n
                        2. I have metadata available for my dataset<\/a>\n
                            \n
                          1. Metadata integration<\/a><\/li>\n\n\n\n
                          2. Subset merged data<\/a><\/li>\n\n\n\n
                          3. UMAP dimensionality reduction of cell cluster abundances<\/a><\/li>\n\n\n\n
                          4. Remove outliers (optional)<\/a><\/li>\n\n\n\n
                          5. Include final UMAP embeddings to the dataset<\/a><\/li>\n\n\n\n
                          6. Hierarchical clustering of UMAP projection<\/a><\/li>\n\n\n\n
                          7. Prepare data for export<\/a><\/li>\n\n\n\n
                          8. Export merged data, boxplots and feature tables<\/a><\/li>\n\n\n\n
                          9. ROC analysis<\/a><\/li>\n<\/ol>\n<\/li>\n\n\n\n
                          10. I do not have metadata available for my dataset<\/a><\/li>\n<\/ol>\n<\/li>\n\n\n\n
                          11. Parameters export<\/a><\/li>\n\n\n\n
                          12. Acknowledgements<\/a><\/li>\n\n\n\n
                          13. Citation<\/a><\/li>\n<\/ol>\n\n\n\n

                            1) Prerequisites<\/strong><\/h2>\n\n\n\n

                            1.1) PICAFlow R package installation<\/strong><\/h3>\n\n\n\n

                            PICAFlow<\/code> package can be installed with the following commands:<\/p>\n\n\n\n

                            # The following line can be skipped if the devtools package is already installed\n<\/em>\ninstall.packages(<\/strong>\"devtools\")<\/strong>\n\n# Load the devtools package\n<\/em>\nlibrary(<\/strong>\"devtools\")<\/strong>\n\n# Install\/update PICAFlow from GitHub repository\n<\/em>\ndevtools::install_github(<\/strong>\"PaulRegnier\/PICAFlow\", force = TRUE)<\/strong><\/code><\/pre>\n\n\n\n
                            1.2) Troubleshooting<\/strong><\/h3>\n\n\n\n
                            In the case the installation of PICAFlow<\/code> is not successful, please check the following points:<\/p>\n\n\n\n