1、Seurat RunTSNE 参数设置
pacman::p_load(Seurat,SeuratData,dplyr,ggplot2)
data("pbmc3k")
cur_seu <- pbmc3k %>% SCTransform() %>% RunPCA() %>% FindNeighbors(dims = 1:25) %>% FindClusters(resolution = 0.3)
cur_seu <- cur_seu %>% RunTSNE(dims = 1:25,
perplexity = 40, ### 参数值越大使得相同簇的空间分布更紧凑,有利于还原全局结构
exaggeration = 0.1) ### 参数值越小使得全局分布更松散,视觉上更圆
DimPlot(cur_seu,reduction = 'tsne')
网友评论