【移轉】CellTypist 註釋免疫細胞亞群

安裝 CellTypist#

配置環境

conda create -n celltypist -c conda-forge r-base=4.1.2
conda activate celltypist
conda install -c conda-forge r-seurat=4.1.0 -y
conda install -c conda-forge r-irkernel=1.3 -y
IRkernel::installspec(name='celltypist', displayname='r-celltypist')
conda install -c conda-forge scanpy=1.8.2 -y
/opt/conda/envs/celltypist/bin/pip3 install celltypist -i https://pypi.tuna.tsinghua.edu.cn/simple
conda install -c conda-forge r-reticulate=1.24 -y
python3
import celltypist
celltypist.models.download_models(force_update = False)

載入套件#

Sys.setenv(RETICULATE_PYTHON = "/opt/conda/envs/celltypist/bin/python3.8")
library(reticulate)
scanpy = import("scanpy")
celltypist = import("celltypist")
pandas <- import("pandas")
numpy = import("numpy")
py_config()

載入資料#

library(Seurat)
sce <- readRDS("~/upload/yy_zhang_data/scRNA-seq/pca.celltype.rds")
Myeloid <- subset(sce, cell_type=='Myeloid')

seurat 轉 scanpy#

# 資料矩陣, scanpy與Seurat的行列定義相反
adata.X = numpy$array(t(as.matrix(Myeloid[['RNA']]@counts)))
# 對每個細胞的觀察
adata.obs = pandas$DataFrame([email protected][colnames(Myeloid[['RNA']]@counts),])
# 對基因矩陣的註釋
adata.var = pandas$DataFrame(data.frame(gene = rownames(Myeloid[['RNA']]@counts), row.names = rownames(Myeloid[['RNA']]@counts)))
 
# 組裝AnnData物件
adata = scanpy$AnnData(X = adata.X, obs=adata.obs, var=adata.var)

進行預測#

model = celltypist$models$Model$load(model = 'Immune_All_AddPIP.pkl')
model$cell_types
scanpy$pp$normalize_total(adata, target_sum=1e4)
scanpy$pp$log1p(adata)
predictions = celltypist$annotate(adata, model = 'Immune_All_AddPIP.pkl', majority_voting = T)

預測添加到 seurat 物件#

Myeloid = AddMetaData(Myeloid, predictions$predicted_labels)