Entering edit mode
9 weeks ago
joonhong kwon ▴ 40
I am analyzing single cell RNA-seq data using Seurat and trying to do trajectory analysis using Monocle3. My analysis pipeline is below.
# QC, NormalizeData, FindVariableFeatures for each sample independetly SelectIntegrationFeatures > FindIntegrationAnchors > IntegrateData > > ScaleData > RunPCA > RunUMAP > FindNeighbors > FindClusters
After then, Sub-clustering and Monocle3 were performed
Oligo <- subset(x, idents=c("OPC", "Oligo1", "Oligo2", "Oligo3") > RunPCA > RunUMAP > FindNeighbors > FindClusters > > as.cell_data_set(Oligo) > ~~ > learn_graph > order_cells
Is this the correct way? Or should other processes such as NormalizeData or FindVariableFeatures be performed after subset?
And when sub-clustering (or re-clustering) is performed, the number of clusters is larger than before. Is it correct to perform trajectory analysis after sub-clustering?