Trajectory inference using PAGA (Scanpy Toolkit)-h5dFormat

1. load libraries

2. Load Data into Seurat

 All_samples_Merged <- readRDS("../0-Seurat_RDS_OBJECT_FINAL/All_samples_Merged_with_STCAT_and_cleaned.rds")


gc()

3. Export to .h5ad (AnnData format)

library(Seurat)
library(SingleCellExperiment)
library(zellkonverter)

# 1️⃣ Set RNA assay as default
DefaultAssay(All_samples_Merged) <- "RNA"

# 2️⃣ Log-normalize counts (creates 'data' slot)
All_samples_Merged <- NormalizeData(
  All_samples_Merged,
  normalization.method = "LogNormalize",
  assay = "RNA",
  verbose = TRUE
)

Normalizing layer: counts
Performing log-normalization
0%   10   20   30   40   50   60   70   80   90   100%
[----|----|----|----|----|----|----|----|----|----|
**************************************************|

# 3️⃣ Extract counts and logcounts
raw_counts <- GetAssayData(All_samples_Merged, slot = "counts", assay = "RNA")

Warning: The `slot` argument of `GetAssayData()` is deprecated as of SeuratObject 5.0.0.
Please use the `layer` argument instead.

logcounts <- GetAssayData(All_samples_Merged, slot = "data", assay = "RNA")

# 4️⃣ Prepare metadata
cell_metadata <- All_samples_Merged[[]]
gene_metadata <- data.frame(
  gene_symbol = rownames(raw_counts),
  row.names = rownames(raw_counts)
)

# 5️⃣ Create SingleCellExperiment (keep sparse matrices)
sce <- SingleCellExperiment(
  assays = list(
    counts = raw_counts, 
    logcounts = logcounts
  ),
  colData = cell_metadata,
  rowData = gene_metadata
)

# 6️⃣ Export to .h5ad
writeH5AD(sce, file = "All_samples_Merged_sparse_logcounts_for_PAGA.h5ad")

ℹ Using the 'counts' assay as the X matrix