更多重要函数见：Seurat重要命令汇总
Seurat绘图函数总结
在使用R语言进行单细胞数据的分析和处理时，除了优秀的绘图包ggplot2以外，Seurat也自带一些优秀的可视化工具，可以用于各种图形绘制。
| Seurat图形绘制函数 | Single gene | Multiple gene | Dimention Reduction | Cluster Information |
|---|---|---|---|---|
| RidgePlot | ✓ | ✓ | ||
| VlnPlot | ✓ | ✓ | ||
| FeaturePlot | ✓ | ✓ | ||
| DotPlot | ✓ | ✓ | ||
| DoHeatmap | ✓ | ✓ | ||
| FeatureScatter | ✓ | ✓ | ||
| Dimplot | ✓ | ✓ |
演示：
示例数据集见monocle中的pbmc3k，并按代码做好注释和保存。
1. RidgePlot山脊图
library(Seurat)
library(tidyverse)
library(patchwork)
rm(list=ls())
dir.create("visual")
pbmc <- readRDS("pbmc.rds")
p1 = RidgePlot(pbmc, features = "FCN1")
p2 = RidgePlot(pbmc, features = "PC_2")
plotc = p1/p2 + plot_layout(guides = 'collect')
ggsave('ridgeplot.png', plotc, width = 10,height = 8)
在山脊图中，features参数设置绘制的对象，可以是基因（查看表达情况），矩阵，PC评分等等可以被FetchData()取出的对象。
上图中，纵轴是细胞类型，原因是该演示数据集已注释好，且细胞类型注释被加入到active.ident中，因此在进行设置时，默认绘制对各个细胞类型进行绘制。可以使用group.by参数进行更改。
p1 = RidgePlot(pbmc, features = "FCN1",group.by = 'seurat_clusters')
p2 = RidgePlot(pbmc, features = "PC_2",group.by = 'seurat_clusters')
plotc = p1/p2 + plot_layout(guides = 'collect')
ggsave('ridgeplot2.png', plotc, width = 10,height = 8)
2. VlnPlot
p1 = VlnPlot(pbmc, features = "nCount_RNA", pt.size = 0,group.by = 'seurat_clusters')
p2 = VlnPlot(pbmc, features = "CD8A", pt.size = 0)
plotc = p1/p2 + plot_layout(guides = 'collect')
ggsave('vlnplot.png', plotc, width = 8,height = 8)
3. FeaturePlot
p1 <- FeaturePlot(pbmc,features = "CD8A", reduction = 'umap')
p2 <- FeaturePlot(pbmc,features = "CCL2", reduction = 'pca')
plotc = p1|p2
ggsave('featureplot.png', plotc, width = 10, height = 4)
对颜色阈值进行设置
p1=FeaturePlot(pbmc, features = "MS4A1")
p2=FeaturePlot(pbmc, features = "MS4A1", min.cutoff = 1, max.cutoff = 3)
p1|p2
在一张图里查看两种marker
FeaturePlot(pbmc, features = c("MS4A1", "CD79A"), blend = TRUE)
4. DotPlot
genelist = c('LYZ','CD79A','CD8A','CD8B','GZMB','FCGR3A')
p = DotPlot(pbmc, features = genelist)
ggsave('dotplot.png', p, width =8, height = 5)
5. DoHeatmap
对聚类分析后鉴定到的每一个分群后进行差异基因鉴定后，每个群选择可以top10的差异基因做的热图展示（热图画的是scale后的@scale.data）。
6. FeatureScatter
p1 <- FeatureScatter(pbmc, feature1 = 'PC_1', feature2 = 'PC_2')
p2 <- FeatureScatter(pbmc, feature1 = 'nCount_RNA', feature2 = 'nFeature_RNA')
plotc = p1|p2
ggsave('featurescatter.png', plotc, width = 10, height = 4)
FeatureScatter还可以用来查看所有细胞或某个细胞中两个基因的相关性
p1=FeatureScatter(pbmc, feature1 = "LYZ", feature2 = "CCL5")
p2=FeatureScatter(pbmc, feature1 = "S100A8", feature2 = "S100A9")
p1|p2
7. Dimplot
p1 <- DimPlot(pbmc, reduction = 'pca', group.by = "cell_type", label=T)
p2 <- DimPlot(pbmc, reduction = 'umap', group.by = "seurat_clusters", label=T)
plotc = p1|p2
ggsave('dimplot.png', plotc, width = 10, height = 4)
8. 交互式绘图
先查看一下所有细胞的注释
DimPlot(pbmc)
如果对某一群细胞感兴趣，可以直接圈出来
select.cells <- CellSelector(plot = plot)
直接手动圈出来想要的细胞，再点Done
head(select.cells)
# [1] "AAACCGTGTATGCG-1" "AAATTCGATTCTCA-1" "AAATTGACACGACT-1"
# [4] "AACCTTACGCGAGA-1" "AACGCCCTCGTACA-1" "AACGTCGAGTATCG-1"
Idents(pbmc, cells = select.cells) <- "NewCells"
# 如果这一群细胞和CD14+ Mono细胞区分不开，就可以找这群细胞和CD14+ Mono相比的特异性marker了
newcells.markers <- FindMarkers(pbmc, ident.1 = "NewCells", ident.2 = "CD14+ Mono", min.diff.pct = 0.3,
only.pos = TRUE)
head(newcells.markers)
参考：https://satijalab.org/seurat/archive/v3.2/visualization_vignette.html