Calculate cell type scores and assign cell types

library(hitype)

Using RunHitype on Seurat object

Prepare the seurat object:

See also https://satijalab.org/seurat/articles/pbmc3k_tutorial.html#setup-the-seurat-object

suppressWarnings(SeuratData::InstallData("pbmc3k"))
pbmc <- pbmc3k.SeuratData::pbmc3k
pbmc <- Seurat::UpdateSeuratObject(pbmc)
pbmc[["percent.mt"]] <- Seurat::PercentageFeatureSet(pbmc, pattern = "^MT-")
pbmc <- subset(pbmc, subset = nFeature_RNA > 200 & nFeature_RNA < 2500 & percent.mt < 5)
pbmc <- Seurat::NormalizeData(pbmc)
pbmc <- Seurat::FindVariableFeatures(pbmc, selection.method = "vst", nfeatures = 2000)
pbmc <- Seurat::ScaleData(pbmc, features = rownames(pbmc))
pbmc <- Seurat::RunPCA(pbmc, features = Seurat::VariableFeatures(object = pbmc))
pbmc <- Seurat::FindNeighbors(pbmc, dims = 1:10)
pbmc <- Seurat::FindClusters(pbmc, resolution = 0.5)
#> Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck
#> 
#> Number of nodes: 2638
#> Number of edges: 95927
#> 
#> Running Louvain algorithm...
#> Maximum modularity in 10 random starts: 0.8728
#> Number of communities: 9
#> Elapsed time: 0 seconds
pbmc <- Seurat::RunUMAP(pbmc, dims = 1:10)

library(hitype)

markers <- data.frame(
    cellName = c(
        "Naive CD4+ T", "CD14+ Mono", "Memory CD4+", "B",
        "CD8+ T", "FCFR3A+ Mono", "NK", "DC", "Platelet"
    ),
    geneSymbolmore1 = c(
        "IL7R,CCR7",  "CD14,LYZ", "IL7R,S100A4", "MS4A1",
        "CD8A", "FCGR3A,MS4A7", "GNLY,NKG7", "FCER1A,CST3", "PPBP"
    ),
    geneSymbolmore2 = rep("", 9)
)

# Load gene sets
gs <- gs_prepare(markers)

# Assign cell types
obj <- RunHitype(pbmc, gs)

Seurat::DimPlot(obj, group.by = "hitype", label = TRUE, label.box = TRUE) +
  Seurat::NoLegend()

Compared to the manual marked cell types:

Seurat manual marked cell types

See also https://satijalab.org/seurat/articles/pbmc3k_tutorial.html#assigning-cell-type-identity-to-clusters

Using hitype_score and hitype_assign on Seurat directly

By default, hitype_score() expects log-normalized data (scaled = FALSE) and performs its own z-scoring. In Seurat v5, access the normalized data with Seurat::GetAssayData(pbmc, layer = "data"):

scores <- hitype_score(Seurat::GetAssayData(pbmc, layer = "data"), gs)
cell_types <- hitype_assign(pbmc$seurat_clusters, scores, gs)
summary(cell_types)
#> # A tibble: 9 × 5
#>   Level Cluster CellType      Score  Margin
#>   <int> <fct>   <chr>         <dbl>   <dbl>
#> 1     1 0       Naive CD4+ T  0.721  0.685 
#> 2     1 1       CD14+ Mono    2.28   1.20  
#> 3     1 2       Memory CD4+   0.599  0.0927
#> 4     1 3       B             2.09   2.17  
#> 5     1 4       NK            1.43   0.179 
#> 6     1 5       FCFR3A+ Mono  3.42   2.44  
#> 7     1 6       NK            3.91   2.88  
#> 8     1 7       DC            5.55   4.59  
#> 9     1 8       Platelet     11.5   11.3

If you prefer to pass pre-scaled data (e.g. from a Seurat scale.data layer), use scaled = TRUE:

scores <- hitype_score(Seurat::GetAssayData(pbmc, layer = "scale.data"),
                       gs, scaled = TRUE)

When scoring with learned weights (see vignette("train-marker-weights")), use norm = "weight", use_sensitivity = FALSE so that scores are normalized by the total marker weight and shared markers are not double-penalized:

scores <- hitype_score(Seurat::GetAssayData(pbmc, layer = "data"), gs,
                       norm = "weight", use_sensitivity = FALSE)

compare to the manual marked cell types:

Cluster ID	Markers	Cell Type
0	IL7R, CCR7	Naive CD4+ T
1	CD14, LYZ	CD14+ Mono
2	IL7R, S100A4	Memory CD4+
3	MS4A1	B
4	CD8A	CD8+ T
5	FCGR3A, MS4A7	FCGR3A+ Mono
6	GNLY, NKG7	NK
7	FCER1A, CST3	DC
8	PPBP	Platelet

See: https://satijalab.org/seurat/articles/pbmc3k_tutorial.html#assigning-cell-type-identity-to-clusters

Exploring the result of hitype_assign

The result of hitype_assign is a data.frame with the following columns:

• Level: the level of the cell type in the hierarchy
• Cluster: the cluster ID
• CellType: the cell type name
• Score: the score of the cell type

head(cell_types)
#> # A tibble: 1 × 5
#>   Level Cluster CellType     Score Margin
#>   <int> <fct>   <chr>        <dbl>  <dbl>
#> 1     1 0       Naive CD4+ T 0.721  0.685