Monocle2: newCellDataSet()

用来生成 Monocle2 中的数据对象 CellDataSet(CDS/cds)

创建一个 CellDataSet

需要：UMI matrix(sparse)+Cell metadata object + Feature metadata object

newCellDataSet(cellData=, phenoData=, featureData=, expressionFamily=)

cellData=: UMI matrix

phenoData=: Cell metadata object, created based on @meta.data in Seurat object

featureData=: Feature metadata object, mainly includes gene ID and gene short name

expressionFamily=: VGAM::negbinomial.size()/VGAM::negbinomial()/VGAM::tobit()/VGAM::gaussianff()

每种数据适合的模型：http://cole-trapnell-lab.github.io/monocle-release/docs/

require(monocle)
require(Seurat)
require(dplyr)

#--------------------------------------------------------#
Seurat_object <- Human_REP1_RPL35_KD_25uM_3_200_filtered_feature_over_2500_HVF_1000_dim_5_G1_phase

#--------------------------------------------------------#
#Extract UMI matrix
UMI_matrix <- as(as.matrix(GetAssayData(Seurat_object, slot = "counts")), "sparseMatrix")

#Extract cell metadata
Seurat_cell_metadata <- Seurat_object@meta.data
Monocle2_cell_metadata <- new("AnnotatedDataFrame", data =Seurat_cell_metadata)   #Create Monocle2 cell metadata object

#Extract gene metadata
Seurat_gene_metadata <- data.frame(gene_id=rownames(Seurat_object), gene_short_name=rownames(Seurat_object))
rownames(Seurat_gene_metadata)<-rownames(Seurat_object)
Monocle2_gene_metadata <- new("AnnotatedDataFrame", data = Seurat_gene_metadata)   #Create Monocle2 gene metadata object

#Create Monocle2 object: CellDataSet(CDS/cds)
Monocle2_object <- newCellDataSet(cellData=UMI_matrix, phenoData = Monocle2_cell_metadata, featureData = Monocle2_gene_metadata, expressionFamily=VGAM::negbinomial.size())

#--------------------------------------------------------#
#Estimate size factor for every cell and dispersion for every feature
Monocle2_object <- estimateSizeFactors(Monocle2_object)   #Estimate size factor for every cell
Monocle2_object <- estimateDispersions(Monocle2_object)   #Estimate dispersion for every feature