5.ArchR的细胞类型定义（2）

愿武艺晴小朋友一定得每天都开心 

---

就为了更好的洞察自己的和ArchR作者究竟在这张UMAP上差多少？

分2步把：1)找到作者一对一颜色的绝对映射，好观察与作者的差别。

在这之前，先去解决掉Rstudio 颜色预览直接显示的问题：这个是跟Rstudio 的版本对应的，对于共有资源Rstudio，俺好像无能为力了，那就这样吧，放过自己。

##我们发现scATAC-seq里面颜色映射的是cluster;进一步去找出作者cluster的具体位置：

#根据作者ATAC中Cluster的颜色，来调整我们PeakMatrix得到的Clusters颜色，尽量逼近作者的
pal <- c("C1"="#8ED2D0","C3"="#FDDD03","C4"="#D51F26","C2"="#5A2955","C5"="#5A2955","C6"="#5A2955","C24"="#FDDD03","C26"="#FDDD03","C25"="#AA875A","C23"="#DB7D8D","C27"="#DB7D8D","C30"="#3E5D8D","C28"="#3E5D8D","C29"="#3EB3A7","C21"="#9C82BA","C22"="#9C82BA", "C20"="#B36B45","C11"="#216069","C10"="#711E21","C9"="#C0545B","C8"="#C0545B","C7"="#C0545B","C17"="#C39C6F","C12"="#245359","C13"="#B4B883","C14"="#F69421","C15"="#753A80","C19"="#753A80","C18"="#2B7F4A","C16"="#DCABCF")
plotEmbedding(
  projHeme2, 
  pal = pal,alpha=0.5, #调整颜色
  labelAsFactors=F, labelMeans=F, #让图显得干净，什么也不label
  size = 0.5,#点的大小，默认是0.1
  embedding = "UMAP-PeakMatrix", 
  name = 'PeakMatrix_Clusters')+labs(title="scATAC-seq(n=35138 cells)")+
  theme_ArchR() + theme(legend.position = "None") 

 ##如何在这张图label 上约束整合最终版的"Celltype"类型？

> cM <- confusionMatrix(projHeme2$PeakMatrix_Clusters, projHeme2$predictedGroup_Co)
> labelOld <- rownames(cM)
> labelOld
 [1] "C27" "C21" "C24" "C25" "C3"  "C4"  "C18" "C15" "C12" "C1"  "C10" "C19" "C17" "C16" "C8"  &#