10X Visium HD 空间基因表达载玻片包含两个 6.5 x 6.5 mm 捕获区域，其中寡核苷酸连续排列在数百万个 2 x 2μm 条形码正方形中，无间隙。数据输出为 2/8/16μm等多个 Bin 尺寸。8x8μm 的 Bin 是可视化和分析的推荐起点。

数据下载：

Visium_HD_Human_Colon_Cancer - Gene expression library of Colon Cancer (Visium HD) using the Human Whole Transcriptome Probe Set

链接：https://www.10xgenomics.com/cn/datasets/visium-hd-cytassist-gene-expression-libraries-of-human-crc-v4

下载以下三个文件：Binned outputs (all bin levels)，Features slice H5，Spatial outputs，并解压spatial.tar.gz与binned_outputs.tar.gz文件。

可以看到结构与Visium基本一致，Binned outputs文件包含：2/8/16μm 三种分辨率.

下面进行数据读取与分析

## Visium HD
#BiocManager::install('arrow')
library(arrow)
library(Seurat)
library(ggplot2)
library(patchwork)
library(dplyr)

##加载 Visium HD 数据
localdir <- "/mnt/My_disk/zhao/DATA/Visium/hd/"

#加载16um数据,减少运行时间
CRC_HD <- Load10X_Spatial(data.dir = localdir, bin.size = c(16))
CRC_HD


#查看数据质量
p1<-VlnPlot(CRC_HD, features = "nCount_Spatial.016um", pt.size = 0) + theme(axis.text = element_text(size = 4)) + NoLegend()

p2<-SpatialFeaturePlot(CRC_HD, features = "nCount_Spatial.016um")

p1|p2

 

#Normalize
CRC_HD <- NormalizeData(CRC_HD)
CRC_HD <- FindVariableFeatures(CRC_HD)
CRC_HD <- ScaleData(CRC_HD)

#select 5w cells and create a 'sketch' assay
CRC_HD <- SketchData(
  object = CRC_HD,
  ncells = 50000,
  method = "LeverageScore",
  sketched.assay = "sketch",
  features = VariableFeatures(CRC_HD)
)

#Switch sketch assay
DefaultAssay(CRC_HD) <- "sketch"

#Clustering workflow
CRC_HD <- FindVariableFeatures(CRC_HD)
CRC_HD <- ScaleData(CRC_HD)
CRC_HD <- RunPCA(CRC_HD, assay = "sketch", reduction.name = "pca.sketch")
CRC_HD <- FindNeighbors(CRC_HD, assay = "sketch", reduction = "pca.sketch", dims = 1:50)
CRC_HD <- FindClusters(CRC_HD, cluster.name = "seurat_cluster.sketched", resolution = 3)
CRC_HD <- RunUMAP(CRC_HD, reduction = "pca.sketch", reduction.name = "umap.sketch", return.model = T, dims = 1:50)

#将从5w个cell中学到的聚类标签和降维投影到整个数据集
CRC_HD <- ProjectData(
  object = CRC_HD,
  assay = "Spatial.016um",
  full.reduction = "full.pca.sketch",
  sketched.assay = "sketch",
  sketched.reduction = "pca.sketch",
  umap.model = "umap.sketch",
  dims = 1:50,
  refdata = list(seurat_cluster.projected = "seurat_cluster.sketched")
)

# switch to sketch assay
DefaultAssay(CRC_HD) <- "sketch"
Idents(CRC_HD) <- "seurat_cluster.sketched"
p1 <- DimPlot(CRC_HD, reduction = "umap.sketch", label = F) + ggtitle("Sketched clustering (50,000 cells)") + theme(legend.position = "bottom")

# switch to full dataset assay
DefaultAssay(CRC_HD) <- "Spatial.016um"
Idents(CRC_HD) <- "seurat_cluster.projected"
p2 <- DimPlot(CRC_HD, reduction = "full.umap.sketch", label = F) + ggtitle("Projected clustering (full dataset)") + theme(legend.position = "bottom")

p1 | p2

 

#空间聚类可视化
SpatialPlot(CRC_HD, pt.size.factor = 6,alpha = 0.8)

 

#指定cluster空间可视化
SpatialPlot(CRC_HD, 
            cells.highlight = CellsByIdentities(object = CRC_HD, idents = c(1, 4, 8)),
            facet.highlight = TRUE, 
            pt.size.factor = 6,
            alpha = 0.8,
            ncol = 3)

 

#Marker可视化
SpatialPlot(CRC_HD, 
            features = c('CD3D','EPCAM','MYH11'),
            pt.size.factor = 6,
            alpha = 0.8,
            ncol = 3)

#保存，下次使用
save(CRC_HD,file = 'CRC_HD.rdata')

 

 

以上就是本期10X Visium HD CRC数据的初步分析，后续分析内容下期见！

更多分析内容，关注 '空间组学' 公众号。