相关性分析 addmodulescore得到的每个细胞评分与page_enrichments得到的每个细胞评分的相关性

相关性分析

addmodule_ns7d_dataframe与PAGE_enrichments相关性

addmodulescore得到的每个细胞评分与page_enrichments得到的每个细胞评分的相关性

library(patchwork)
library(ggplot2)
library(ggalluvial)
library(svglite)
library(Seurat)
library(openxlsx)
library(Hmisc)
#https://www.jianshu.com/p/cef5663888ff
getwd()
path="G:/silicosis/sicosis/silicosis_ST/overlapped_map/addmodule_ns7d"
dir.create(path)
setwd(path)

#load("G:/silicosis/sicosis/silicosis_ST/yll/0214/harmony_cluster/d_all/silicosis_ST_harmony_SCT_r0.6.rds")
load("G:/silicosis/需求/矽肺-数据分析结果-0119-yll/矽肺-数据分析结果-0119/single_slide_cluster/silicosis_ST_NS_7_SCT_r0.5.rds")
table(Idents(NS_7_sct))  # 0   1   2   3

#marker = read.xlsx("G:/silicosis/sicosis/silicosis_ST/overlapped_map/Rigional and cell markers.xlsx", 
#                   sheet = "SingleCell_markers")
marker = read.xlsx("G:/silicosis/sicosis/gitto/myspatial/SingCellMarkers.xlsx",
                             sheet='selected_markers')

head(marker)
library(dplyr)

marker=marker[,-1]

cellnames=colnames(marker)                ##number=length(marker[,cellname])



for (each in cellnames) {
  # each="macroph-1"
  cellname=each
  
  mymarker=marker[,cellname]   %>% na.exclude() %>% unique() %>%
    capitalize()    %>% list()
  number=length(mymarker[[1]])
  unlist(mymarker)
  
  #对给定的基因集合进行打分  并画图
  if(1==1){NS_7_sct=AddModuleScore(NS_7_sct,
                                   features = mymarker,
                                   name = paste0(cellname,"_Addmodule"))
  #结果保存在这里
  colnames(NS_7_sct@meta.data)
  
   }
}
colnames(NS_7_sct@meta.data)
#NS_7_sct
mydata=FetchData(NS_7_sct,vars = colnames(NS_7_sct@meta.data))
head(mydata)


data=mydata[,-c(1:8)]
head(data)
getwd()
#save(data,file = "G:/silicosis/sicosis/silicosis_ST/overlapped_map/addmodule_ns7d/addmodule_ns7d_dataframe.rds")
load("G:/silicosis/sicosis/silicosis_ST/overlapped_map/addmodule_ns7d/addmodule_ns7d_dataframe.rds")



##得到page_enrichments结果的代码
if(1==1){library(svglite)
  library(Seurat)
  library(xlsx)
  library(Giotto)
  #remotes::install_github("RubD/Giotto", build_vignettes = T)
  
  load("G:/silicosis/需求/矽肺-数据分析结果-0119-yll/矽肺-数据分析结果-0119/single_slide_cluster/silicosis_ST_NS_7_SCT_r0.5.rds")
  table(Idents(NS_7_sct))  # 0   1   2   3
  
  load("G:/silicosis/需求/矽肺-数据分析结果-0119-yll/矽肺-数据分析结果-0119/single_slide_cluster/silicosis_ST_SiO2_7_SCT_r0.6.rds")
  Idents(sio2_7_sct) # 0   1   2   3   4   5 6
  
  
  load("G:/silicosis/需求/矽肺-数据分析结果-0119-yll/矽肺-数据分析结果-0119/single_slide_cluster/silicosis_ST_NS_56_SCT_r0.5.rds")
  table(Idents(NS_56_sct)) #0   1   2   3
  
  load(file = "G:/silicosis/需求/矽肺-数据分析结果-0119-yll/矽肺-数据分析结果-0119/single_slide_cluster/silicosis_ST_SiO2_56_SCT_r0.5_true.rds")
  table(Idents(sio2_56_sct))  # 0   1   2   3   4   5 
  
  
  ##ns_7d
  getwd()
  path="G:/silicosis/sicosis/gitto/seurat/ns7d-2"
  dir.create(path,recursive = TRUE)
  setwd(path)
  
  load("G:/silicosis/需求/矽肺-数据分析结果-0119-yll/矽肺-数据分析结果-0119/single_slide_cluster/silicosis_ST_NS_7_SCT_r0.5.rds")
  table(Idents(NS_7_sct))  # 0   1   2   3
  
  DefaultAssay(NS_7_sct)
  str(NS_7_sct)
  
  
  #load("G:/silicosis/需求/矽肺-数据分析结果-0119-yll/矽肺-数据分析结果-0119/NS_7.rds")
  
  #seurat to giotto  自建函数转化 
  seuratToGiotto_OLD <- function(obj_use = NULL,...){
    require(Seurat)
    require(Giotto)
    #obj_use=NS_7_sct
    # get general info in basic seurat structures
    obj_assays <- names(obj_use@assays)
    if ('Spatial' %in% obj_assays){
      obj_assays <- c('Spatial',obj_assays[-which(obj_assays == 'Spatial')])
    }
    
    obj_dimReduc <- names(obj_use@reductions)
    obj_dimReduc_assay <- sapply(obj_dimReduc,function(x) 
      #x="pca"
      obj_use[[x]]@assay.used)
    
    obj_graph_expr <- names(obj_use@graphs) #"SCT_nn"  "SCT_snn"
    obj_graph_expr_assay <- sapply(obj_graph_expr,function(x) 
      #x="SCT_nn"
      obj_use[[x]]@assay.used)
    
    obj_meta_cells <- obj_use@meta.data #head(obj_use@meta.data) SCT_snn_res.0.4 seurat_clusters orig.ident nCount_Spatial nFeature_Spatial stim nCount_SCT nFeature_SCT
    obj_meta_genes <- lapply(obj_assays,function(x) 
      obj_use[[x]]@meta.features)
    names(obj_meta_genes) <- obj_assays
    
    obj_img <- obj_use@images
    obj_img_names <- names(obj_img) #"image"
    loc_use <- lapply(obj_img_names,function(x){
      temp <- obj_img[[x]]@coordinates
      temp <- as.data.frame(temp[,c('col','row')])
      # temp$region <- x
      return (temp)
    })
    loc_use <- Reduce(rbind,loc_use)
    
    # add assay data: raw, normalized & scaled
    for (i in c(2) ){ #1:length(obj_assays)
      #i=2
      data_raw <- GetAssayData(obj_use,slot = 'counts',assay = obj_assays[i])
      data_norm <- GetAssayData(obj_use,slot = 'data',assay = obj_assays[i])
      data_scale <- GetAssayData(obj_use,slot = 'scale.data',assay = obj_assays[i])
      #head(data_scale)
      #DefaultAssay(obj_assays)
      #DefaultAssay(obj_use)
      #head(obj_use@assays$SCT@scale.data)
      #table(obj_use@assays$SCT@scale.data>0)
      if (1==1){ #i == 1 & obj_assays[i] == 'Spatial'
        feat_use <- 'rna'
        test <- createGiottoObject(raw_exprs = obj_use[[obj_assays[i]]]@counts,
                                   norm_expr=data_norm,
                                   #norm_scaled_expr =data_scale,
                                   spatial_locs = loc_use #slotName = 'rna'
        )
        test <- addCellMetadata(test,new_metadata = obj_meta_cells) #,feat_type = feat_use
      } else {
        #i=2
        feat_use <- obj_assays[i]
        test@expression[[feat_use]][['raw']] <- data_raw
        test@feat_ID[[feat_use]] = rownames(data_raw)
        test@feat_metadata[[feat_use]] = data.table::data.table(feat_ID = test@feat_ID[[feat_use]])
      }
      
      
      
      
      # add dim reduction
      for (i in obj_dimReduc){
        if (!i %in% c('pca','umap','tsne')){
          next
        } else {
          dimReduc_name <- i
          dimReduc_method <- i
          dimReduc_coords <- obj_use[[i]]@cell.embeddings
          dimReduc_misc <- list(obj_use[[i]]@stdev,
                                obj_use[[i]]@feature.loadings,
                                obj_use[[i]]@feature.loadings.projected)
          names(dimReduc_misc) <- c('eigenvalues','loadings','loadings_projected')
          dimObject <- Giotto:::create_dimObject(name = dimReduc_name,reduction_method = dimReduc_method,
                                                 coordinates = dimReduc_coords,misc = dimReduc_misc)
          test@dimension_reduction[['cells']][[dimReduc_method]][[dimReduc_name]] <- dimObject
        }
      }
      
      # add expr nearest neighbors
      for (i in obj_graph_expr){
        mtx_use <- obj_use[[i]]
        ig_use <- igraph::graph_from_adjacency_matrix(mtx_use,weighted = T)
        g_type <- unlist(strsplit(i,split = "_"))[2]
        g_val <- i
        test@nn_network[[g_type]][[g_val]][['igraph']] <- ig_use
      }
      return (test)
      
    }
  }
  
  Assays(NS_7_sct@assays$SCT)
  NS_7_sct@assays$SCT@scale.data
  NS_7_sct_giotto=seuratToGiotto_OLD(NS_7_sct)
  
  ##########part 7: cell-type annotation######################################################
  # known markers for different mouse brain cell types:
  # Zeisel, A. et al. Molecular Architecture of the Mouse Nervous System. Cell 174, 999-1014.e22 (2018).
  ## cell type signatures ##
  ## combination of all marker genes identified in Zeisel et al
  #sign_matrix_path = system.file("extdata", "sig_matrix.txt", package = 'Giotto')
  getwd()
  library(openxlsx)
  library(Hmisc)
  brain_sc_markers = read.xlsx("G:/silicosis/sicosis/gitto/myspatial/SingCellMarkers.xlsx",
                               sheet='Sheet2')
  head(brain_sc_markers)
  topper_column=capitalize(colnames(brain_sc_markers))
  colnames(brain_sc_markers)=topper_column
  
  topper_markers=capitalize(brain_sc_markers$Event)
  
  head(topper_markers)
  brain_sc_markers$Event=topper_markers
  
  head(brain_sc_markers)
  sig_matrix = as.matrix(brain_sc_markers[,-1]); rownames(sig_matrix) = brain_sc_markers$Event
  sig_matrix[is.na(sig_matrix)] <- 0
  head(sig_matrix)
  
  
  visium_brain=NS_7_sct_giotto
  
  
  
  
  
  ## enrichment tests 
  visium_brain = createSpatialEnrich(visium_brain, 
                                     sign_matrix = sig_matrix, 
                                     enrich_method = 'PAGE',
                                     min_overlap_genes = 1)
  visium_brain = runPAGEEnrich(gobject = visium_brain, 
                               sign_matrix = sig_matrix,
                               min_overlap_genes = 0) #可以改变min_overlap_genes的数量
  
  getwd()
  ## heatmap of enrichment versus annotation (e.g. clustering result)
  cell_types = colnames(sig_matrix)
  head(cell_types)
  length(cell_types)
  plotMetaDataCellsHeatmap(gobject = visium_brain,
                           metadata_cols = 'seurat_clusters',
                           value_cols = cell_types,spat_enr_names = 'PAGE',
                           x_text_size = 8, y_text_size = 8,
                           save_param = list(save_name="7_a-_metaheatmap"))
  
  getwd()
  #save(visium_brain,file ="G:/silicosis/sicosis/gitto/seurat/ns7d-2/ns_7d_sct_giotto.rds" )
  load("G:/silicosis/sicosis/gitto/seurat/ns7d-2/ns_7d_sct_giotto.rds")
  #load("G:/silicosis/sicosis/gitto/seurat/ns7d/NS_7_sct_dwl_after_dwl.rds")
  
  
  str(NS_7_sct_dwl)
  
  getwd()
  page_enrichments=as.data.frame(visium_brain@spatial_enrichment$PAGE)
  head(page_enrichments)
  rownames(page_enrichments)<-page_enrichments$cell_ID
  page_enrichments=page_enrichments[,-1]
  head(page_enrichments)
  class(page_enrichments)
  #save(page_enrichments,file = "G:/silicosis/sicosis/gitto/seurat/ns7d-2/PAGE_enrichments.rds")
}
load("G:/silicosis/sicosis/gitto/seurat/ns7d-2/PAGE_enrichments.rds")



head(data)

head(page_enrichments)

mydata=bind_cols(data,page_enrichments)
head(mydata)

#求相关性
cor_data <- cor(mydata,method="pearson")
cor_data2 <- cor(mydata,method="spearman")
cor_data[1:5,1:5]


# Compute a matrix of correlation p-values
p.mat <- ggcorrplot::cor_pmat(mydata)
head(p.mat[, 1:4])

#可视化     #https://www.jianshu.com/p/aeb9f612e888
library(corrplot)


##默认参数
corrplot(cor_data)
p=corrplot(cor_data,order="hclust",hclust.method="average")#,addrect=5

##使用聚类顺序
ggcorrplot::ggcorrplot(cor_data,hc.order=TRUE,outline.color="white")


p=ggcorrplot::ggcorrplot(cor_data,outline.color="white")

p=ggcorrplot::ggcorrplot(cor_data,outline.color="white",  p.mat = p.mat,
                         hc.order = TRUE,
                         type = "lower",
                         insig = "blank")

p=ggcorrplot::ggcorrplot(cor_data,outline.color="white",  p.mat = p.mat,
                         hc.order = TRUE,
                         insig = "blank")

p=ggcorrplot::ggcorrplot(
  cor_data,
  p.mat = p.mat,
  
  insig = "blank"
)#hc.order = TRUE,
jpeg("correaltaion-67.jpeg",width = 20,height = 20,units = 'in', res=600)
print(p)
dev.off()

##测试用
for (each in cellnames) {
 # each="macroph-1"
  cellname=each
  
  mymarker=marker[,cellname]   %>% na.exclude() %>% unique() %>%
    capitalize()    %>% list()
  number=length(mymarker[[1]])
  unlist(mymarker)
  
  #对给定的基因集合进行打分  并画图
  if(1==1){NS_7_sct=AddModuleScore(NS_7_sct,
                                features = mymarker,
                                name = "mymarker")
  #结果保存在这里
  colnames(d.all@meta.data)
  
  ###
  
  p1=SpatialFeaturePlot(d.all, features = "mymarker1", slot = "scale.data",images = "image")+ ggtitle(paste(unlist(mymarker), collapse = "|"))
  p2=SpatialFeaturePlot(d.all, features = "mymarker1", slot = "scale.data",images = "image.1")
  p3=SpatialFeaturePlot(d.all, features = "mymarker1", slot = "scale.data",images = "image.2")+ ggtitle(paste(cellname))
  p4=SpatialFeaturePlot(d.all, features = "mymarker1", slot = "scale.data",images = "image.3")
  
  jpeg(paste0(cellname,"_","total_",length(unlist(mymarker)),"_",paste0(min(number),"-",max(number)), 
              paste(unlist(mymarker)[1:15],collapse = "_"),"_.jpeg"), #只取前15个
       height = 12, width = 12, units = 'in', res=600)
  p=ggpubr::ggarrange(p1,p2,p3,p4,ncol = 2,nrow =2)
  print(p)
  dev.off()}
}
··

#下面的代码是为了测试
if(1==1){#
  names(d.all)
  
  SpatialFeaturePlot(d.all, features = "mymarker1", slot = "scale.data")+ 
    ggtitle(paste(unlist(mymarker), collapse = "|"))
  
  #https://ggplot2-book.org/polishing.html?q=title#themes
  SpatialFeaturePlot(d.all, features = "mymarker1", slot = "scale.data")+ 
    theme(plot.title =element_text(paste(unlist(mymarker), collapse = "|")) )
  
  
  ####
  SpatialFeaturePlot(d.all, features = "mymarker1", slot = "scale.data")+ 
    labs(title = paste(unlist(mymarker), collapse = "|"),
         colour="Cylinders")
  
}