seurat对象中细胞identity的获取、设置与操纵

前两天遇到了一个小问题：初步注释细胞发现，使用RenameIdents后细胞类型的levels与我想要的排序不符。例如，我实际得到的level为: cancer_cell1 T_cells cancer_cell2 cancer_cell3 B_cells.但是我想要把cancer_cell cluster与免疫细胞的排列成：cancer_cell1 cancer_cell2 cancer_cell3 T_cells B_cells。搜所了一下，结果方法很简单。

# Reorder identity classes

levels(x = pbmc_small)

#> [1] "B" "A" "C"

levels(x = pbmc_small) <- c('C', 'A', 'B')

levels(x = pbmc_small)

#> [1] "C" "A" "B"

seurat处理identity的函数有

Idents(object, ...)

Idents(object, ...) <- value

RenameIdents(object, ...)

ReorderIdent(object, var, ...)

SetIdent(object, ...)

StashIdent(object, save.name, ...)

# S3 method for Seurat

Idents(object, ...)

# S3 method for Seurat

Idents(object, cells = NULL, drop = FALSE, ...) <- value

# S3 method for Seurat

ReorderIdent(

  object,

  var,

  reverse = FALSE,

  afxn = mean,

  reorder.numeric = FALSE,

  ...

)

# S3 method for Seurat

RenameIdents(object, ...)

# S3 method for Seurat

SetIdent(object, cells = NULL, value, ...)

# S3 method for Seurat

StashIdent(object, save.name = "orig.ident", ...)

# S3 method for Seurat

droplevels(x, ...)

# S3 method for Seurat

levels(x)

# S3 method for Seurat

levels(x) <- value

# 获取细胞identy

Idents(pbmc_small)

# 设置细胞identy

# 可以用来指定某些细胞为新的名称

#例如将前4个细胞的identity修改为a，其余不变

Idents(pbmc_small, cells = 1:4) <- 'a'

head(Idents(pbmc_small))

# 使用matadata来设置identity

colnames(pbmc_small[[]])

Idents(pbmc_small) <- 'RNA_snn_res.1'

levels(pbmc_small)

# 修改细胞identity的名字

# 可以提供任意数量，例如，原始的level有20，修改2个，剩余的18个保持不变

levels(pbmc_small)

pbmc_small <- RenameIdents(pbmc_small, '0' = 'A', '2' = 'C')

levels(pbmc_small)

# 这里也可以写成

identityMapping <- c('0' = 'A', '2' = 'C')

pbmc_small <- RenameIdents(pbmc_small, identityMapping )

# 使用SetIdent设置细胞的identity

cells.use <- WhichCells(pbmc_small, idents = '1')

pbmc_small <- SetIdent(pbmc_small, cells = cells.use, value = 'B')

#将细胞的identity加到matadata上

head(pbmc_small[[]])

pbmc_small <- StashIdent(pbmc_small, save.name = 'idents')

head(pbmc_small[[]])

# 获取seurat对象的细胞identity的水平

levels(x = pbmc_small)

# 调整细胞identity的水平

levels(x = pbmc_small)

levels(x = pbmc_small) <- c('C', 'A', 'B')

levels(x = pbmc_small)

感觉上面的函数基本上涵盖了操纵identity的所有需求。

如有错误，可以在评论初指正。

参考

• https://mojaveazure.github.io/seurat-object/reference/Idents.html