CyteTypeR Annotation - PRE-SAVED Clusters-Normal Integrated CD4 T cells

1 load libraries

2 Load Seurat Object

# Load object WITH pre-saved clusters & markers
ref_obj <- readRDS("../CD4_reference_annotated_with_markers.rds")  # ← your saved object

# Verify markers are there
print("Available markers:")
print(names(ref_obj@misc)[grepl("markers", names(ref_obj@misc))])
head(ref_obj$markers_top25)  # per-cell top markers

3 Extract Pre-Computed Markers for CyteTypeR

library(dplyr)

# Use TOP25 markers (recommended for cytypeR robustness)
SS_markers_filtered <- ref_obj@misc$markers_top25  # already filtered + top25/cluster

# Or ALL significant if you prefer:
SS_markers_filtered <- ref_obj@misc$markers_all_filtered

cat("Markers loaded:", nrow(SS_markers_filtered), "\n")

Markers loaded: 5249 

print(head(SS_markers_filtered[, c("cluster", "gene", "avg_log2FC")]))

       cluster   gene avg_log2FC
EEF1B2       0 EEF1B2  0.2696944
CCR7         0   CCR7  0.6658896
LEF1         0   LEF1  0.6232084
TCF7         0   TCF7  0.4312715
FHIT         0   FHIT  1.3244055
TMIGD2       0 TMIGD2  2.2133935

4 Run CyteTypeR with Pre-Saved Data

# Prepare using PRE-SAVED markers (no FindAllMarkers needed!)
prepped_data <- PrepareCyteTypeR(
  obj = ref_obj,
  SS_markers_filtered,
  n_top_genes = 25,
  group_key = 'seurat_clusters',
  aggregate_metadata = TRUE,
  coordinates_key = "umap"
)

# ---------------- Professional CyteTypeR Metadata -----------------------
metadata <- list(
  title = 'Human CD4+ T Cell Subset Atlas - Integrated Reference',
  run_label = 'CD4T_10xGenomics_Lab_Integrated_v1',
  experiment_name = 'Primary CD4 T cells from healthy donors (10x + lab CITE-seq)'
)

# Clean cluster metadata
prepped_data$clusterMetadata <- prepped_data$clusterMetadata[
  !is.na(names(prepped_data$clusterMetadata)) & 
  names(prepped_data$clusterMetadata) != "NA"
]
print("Cleaned cluster metadata:")
print(names(prepped_data$clusterMetadata))

# ---------------- Detailed Study Context -------------------------------
results <- CyteTypeR(
  obj = ref_obj,
  prepped_data = prepped_data,
  study_context = paste0(
    "Unsupervised clustering (0-6) of 11,482 high-quality CD4+ T cells from 3 healthy donors. ",
    "Integrated analysis combining 10x Genomics CITE-seq (CD4T_10x_S1/S2) + lab-generated ",
    "ADT data (CD4T_lab). RPCA+SCT integration with doublet removal. ",
    "Markers derived from RNA assay (top 25/cluster, p_adj<0.05, logFC>0.25, ",
    "min.pct.diff>0.20). Blacklist-filtered (TCR/Ig/ribo/MT/HB). ",
    "Objective: Automated cell-type annotation of naive/memory/effector/Treg subsets."
  ),
  metadata = metadata
)

saveRDS(results, "CyteTypeR_CD4T_integrated_reference.rds")
cat("✅ CyteTypeR results saved with professional metadata\n")

5 Visualize Results

# UMAPs with annotations
p1 <- DimPlot(results, group.by = "cytetype_cell_ontology_seurat_clusters", 
              label = TRUE, label.box = TRUE, repel = TRUE, reduction = "umap")
p2 <- DimPlot(results, group.by = "cytetype_seurat_clusters", reduction = "umap")

p1 | p2

# Compare original vs cytypeR annotation
p3 <- DimPlot(ref_obj, group.by = "seurat_clusters", label = TRUE, label.box = TRUE)
p4 <- DimPlot(results, group.by = "cytetype_seurat_clusters", label = TRUE)

(p3 | p4) / (p1 | p2)

6 Validation Plots (Optional)

## Use top markers from your saved object for validation
top_markers <- ref_obj@misc$markers_top5$gene[1:12]  # first 12 top markers
available <- intersect(top_markers, rownames(ref_obj))

FeaturePlot(ref_obj, features = available, ncol = 4, pt.size = 0.3)

7 Summary Table

# Compare original clusters vs cytypeR predictions
table(
  Original = ref_obj$seurat_clusters,
  CytypeR = results$cytetype_seurat_clusters
)

        CytypeR
Original Activated CD4+ T helper cells Activated effector regulatory T cell
       0                             0                                    0
       1                          3998                                    0
       2                             0                                    0
       3                             0                                    0
       4                             0                                    0
       5                             0                                  341
       6                             0                                    0
        CytypeR
Original CD4+ cytotoxic T lymphocytes CD4+ Th2-polarized effector T cell
       0                            0                                  0
       1                            0                                  0
       2                            0                                522
       3                          491                                  0
       4                            0                                  0
       5                            0                                  0
       6                            0                                  0
        CytypeR
Original Naive CD4+ T cell (IFITM3+) Naive thymic CD4+ T cell (RFX3+)
       0                        5485                                0
       1                           0                                0
       2                           0                                0
       3                           0                                0
       4                           0                                0
       5                           0                                0
       6                           0                              233
        CytypeR
Original Stressed/dying CD4+ T lymphocyte
       0                                0
       1                                0
       2                                0
       3                                0
       4                              412
       5                                0
       6                                0

# Save final annotated object
ref_obj$cytetype_annotation <- results$cytetype_seurat_clusters


#saveRDS(ref_obj, "CD4_reference_cytetype_annotated.rds")

#<https://nygen-labs-prod--cytetype-api.modal.run/report/9d726cf3-bdce-4f96-9610-c3a45a77a611>