Hello everyone

I am working on four samples from 10X spatial transcriptome data. Here data1_f and data2_f belongs to control while data3_f and data4_f belong to inflammed group. I applied following step for the batch correction by Harmony (based on the sample id).

Each sample was analysed for SCTransform -> merged for single seurat object -> RunPCA -> RunHarmony -> RunUMAP

object <- merge(data1_f, y=c(data2_f,data3_f,data4_f), add.cell.ids = c("Donor1","Donor2","Donor3","Donor4"), project = "pbmc_combined")

pancreas.list <- SplitObject(object = object, split.by = "orig.ident" )


for (i in 1:length(pancreas.list)) {
    pancreas.list[[i]] <- SCTransform(pancreas.list[[i]], verbose = FALSE,assay = "Spatial")
}


pancreas.features <- SelectIntegrationFeatures(object.list = pancreas.list, nfeatures = 3000)

c1 <- pancreas.list[[1]]


d1 <- pancreas.list[[2]]


a1 <- pancreas.list[[3]]


b1 <- pancreas.list[[4]]


pancreas_merged <- merge(x=c1, y = c(d1,a1,b1),  project = "pancreas", merge.data = TRUE)


VariableFeatures(pancreas_merged) <- pancreas.features
pancreas_merged <- RunPCA(object = pancreas_merged, assay = "SCT", npcs = 50)

pancreas_merged <- RunHarmony(object = pancreas_merged,
                                    assay.use = "SCT",
                                    reduction = "pca",
                                    dims.use = 1:50,
                                    group.by.vars = "orig.ident",
                                    plot_convergence = TRUE)


pancreas_merged <- RunUMAP(object = pancreas_merged, assay = "SCT", reduction = "harmony", dims = 1:50)

After batch correction, I was expecting overlapped cluster from both the groups. But it give me two distinct cluster based on the condition. I suppose I am doing something wrong during the analysis. I would appreciate all the suggestion.