整合pbmc1k和pbmc3k数据数据基本情况：nFeature_RNAnCount_RNApercent.mt>pbmc1kAnobject ofclass Seurat23415featuresacross1e99sampleswithin4assaysActiveassay:RNA(23148 features,2000 variable features)3 other assays present: prediction.score.celltype.l1,prediction.score.celltype.12,predicted_ADT5 dimensional reductions calculated:pca, umap, tsne,ref.spca,ref.umap300>pbmc3kAn obiect of class.Seurat13981 features across 2638 sampleswithin4 assaysActive assay:RNA (13714 features,2ee variable features)3 other assays present: prediction.score.celltype.l1,prediction.score.celltype.12,predicted_ADT5 dimensional reductions calculated: pca, umap, tsne, ref.spca,200ref.umap>pbmc_objAn object of class Seurat25823 featuresacross 3737sampleswithin4 assaysActive assay:RNA (25556 features, variable features)3 other assays present: prediction.score.celltype.l1,prediction.score.celltype.12,predicted_ADTIder>上节回顾数据整合差异基因富集分析拟时序分析细胞通讯
整合 pbmc1k 和 pbmc3k 数据 上节回顾 数据整合 差异基因富集分析 拟时序分析 细胞通讯 数据基本情况： > pbmc1k An object of class Seurat 23415 features across 1099 samples within 4 assays Active assay: RNA (23148 features, 2000 variable features) 3 other assays present: prediction.score.celltype.l1, prediction.score.celltype.l2, predicted_ADT 5 dimensional reductions calculated: pca, umap, tsne, ref.spca, ref.umap > pbmc3k An object of class Seurat 13981 features across 2638 samples within 4 assays Active assay: RNA (13714 features, 2000 variable features) 3 other assays present: prediction.score.celltype.l1, prediction.score.celltype.l2, predicted_ADT 5 dimensional reductions calculated: pca, umap, tsne, ref.spca, ref.umap > pbmc_obj An object of class Seurat 25823 features across 3737 samples within 4 assays Active assay: RNA (25556 features, 0 variable features) 3 other assays present: prediction.score.celltype.l1, prediction.score.celltype.l2, predicted_ADT
整合pbmc1k和pbmc3k数据不考虑任何批次效应直接整合：存在明显的批次效应orig.identmarker_celltype10.B.CD14+Mono2.CD4spbmc1kCD8T.ptmc3k.DC-FCGR3A+Mono.NKPlatelet--10-1010-10UMAP_1UMAP_1>上节回顾>数据整合差异基因富集分析拟时序分析细胞通讯
整合 pbmc1k 和 pbmc3k 数据 不考虑任何批次效应直接整合：存在明显的批次效应 上节回顾 数据整合 差异基因富集分析 拟时序分析 细胞通讯
利用Seurat的RPCA整合数据pbmc1k和pbmc3k数据示例（k.anchor=5）marker_celltypeorig.ident2010FCGF.CD14+Mono-CD4CD8Tepbmc1k..DCDbme3kFCGR3A+Mono..NKPlstPlatelet-10-10-10010-10010UMAP_1UMAP_1>上节回顾数据整合差异基因富集分析拟时序分析细胞通讯
利用 Seurat 的RPCA整合数据 pbmc1k和pbmc3k数据示例（k.anchor = 5） 上节回顾 数据整合 差异基因富集分析 拟时序分析 细胞通讯
RPCA整合与简单合并在分析上的差异简单合并pbmc_obj<-NormalizeData(pbmc obj,normalization.method ="LogNormalize",scale.factor =1oooo)pbmc obj<-FindVariableFeatures(pbmcobj,selection.method="vst",nfeatures=2ooo)all.genes<-rownames(pbmcobj)pbmc_obj<-ScaleData(pbmc_obj,features =all.genes)pbmc_obj<-RunPCA(pbmc_obj,features=VariableFeatures(object=pbmc_obj))pbmc_obj<-FindNeighbors(pbmc_obj,dims=1:1o)pbmc_obj<-Findclusters(pbmc_obj,resolution=.5)pbmc_obj<-RunUMAP(pbmc_obj,dims=1:1)pbmc_obj<-RunTSNE(pbmc_obj,dims=1:10)RPCA整合#selectfeaturesthatarerepeatedlyvariablefeatures <- SelectIntegrationFeatures(object.list-pbmc)pbmc<-lapply(X=pbmc,FUN=function(x)(acrossdatasetsfor integration run PCAon eachx<-ScaleData(x,features =features,verboseFALsE)dataset using thesefeaturesx<-RunPCA(x,features =features,verbose=FALSE)3)anchors <-FindIntegrationAnchors(object.list -pbmc,anchor.features =features, reduction -"rpca")combined <-IntegrateData(anchorset =anchors)DefaultAssay(combined)<-"integrated">上节回顾>数据整合拟时序分析细胞通讯差异基因富集分析
RPCA整合与简单合并在分析上的差异 上节回顾 数据整合 差异基因富集分析 拟时序分析 细胞通讯 pbmc_obj <- NormalizeData(pbmc_obj, normalization.method = "LogNormalize", scale.factor = 10000) pbmc_obj <- FindVariableFeatures(pbmc_obj, selection.method = "vst", nfeatures = 2000) all.genes <- rownames(pbmc_obj) pbmc_obj <- ScaleData(pbmc_obj, features = all.genes) pbmc_obj <- RunPCA(pbmc_obj, features = VariableFeatures(object = pbmc_obj)) pbmc_obj <- FindNeighbors(pbmc_obj, dims = 1:10) pbmc_obj <- FindClusters(pbmc_obj, resolution = 0.5) pbmc_obj <- RunUMAP(pbmc_obj, dims = 1:10) pbmc_obj <- RunTSNE(pbmc_obj, dims = 1:10) 简单合并 RPCA整合 features <- SelectIntegrationFeatures(object.list = pbmc) pbmc <- lapply(X = pbmc, FUN = function(x) { x <- ScaleData(x, features = features, verbose = FALSE) x <- RunPCA(x, features = features, verbose = FALSE) }) anchors <- FindIntegrationAnchors(object.list = pbmc, anchor.features = features, reduction = "rpca") combined <- IntegrateData(anchorset = anchors) DefaultAssay(combined) <- "integrated" # select features that are repeatedly variable across datasets for integration run PCA on each dataset using these features
利用Seurat的RPCA整合数据pbmc1k和pbmc3k数据示例（FindlntegrationAnchors函数中k.anchor=20来增加整合力度orig.idermarker_celltype10B.CD14+ Mono.CD4CDBTDCFCGR3A+MonoNKPlatelet5101510-510-10-5101UMAP1UMAP_1上节回顾>数据整合差异基因富集分析拟时序分析细胞通讯
利用 Seurat 的RPCA整合数据 pbmc1k和pbmc3k数据示例（FindIntegrationAnchors函数中k.anchor = 20来增加整合力度） 上节回顾 数据整合 差异基因富集分析 拟时序分析 细胞通讯