CARD反卷积笔记

引自https://yingma0107.github.io/CARD/documentation/04_CARD_Example.html

# install the CARD package
devtools::install_github('YingMa0107/CARD')

# load package
library(CARD)

wget https://github.com/YingMa0107/CARD/raw/master/data/spatial_location.RData
wget https://github.com/YingMa0107/CARD/raw/master/data/spatial_count.RData
wget https://github.com/YingMa0107/CARD/raw/master/data/sc_meta.RData
wget https://github.com/YingMa0107/CARD/raw/master/data/sc_count.RData

1.CARD去卷积需要的input数据

1）单细胞sc_count.RData、sc_meta.RData

#### load the example scRNA-seq count data, 
load("./sc-data/sc_count.RData")
sc_count[1:4,1:4]
#The scRNA-seq count data must be in the format of matrix or sparseMatrix, 
#while each row represents a gene and each column represents a cell.
sc_count
#### load the example scRNA-seq meta data, 
load("./sc-data/sc_meta.RData")
sc_meta[1:4,]
#The scRNAseq meta data must be in the format of data frame while each row represents a cell. 
#The rownames of the scRNAseq meta data should match exactly with the column names of the scRNAseq count data. 
#The sc_meta data must contain the column indicating the cell type assignment for each cell (e.g., “cellType” column in the example sc_meta data). 
#Sample/subject information should be provided, if there is only one sample, we can add a column by 
#sc_meta$sampleInfo = "sample1".

#We suggest the users to check their single cell RNASeq data carefully before running CARD. 
#We suggest the users to input the single cell RNAseq data with each cell type containing at least 2 cells. 
#i.e. print(table(sc_meta$cellType,useNA = “ifany”))
sc_meta

2）空转spatial_count.RData、spatial_location.RData

#### load the example spatial transcriptomics count data, 
load("./spa-data/spatial_count.RData")
spatial_count[1:4,1:4]
#The spatial transcriptomics count data must be in the format of matrix or sparseMatrix, 
#while each row represents a gene and each column represents a spatial location. 
#The column names of the spatial data can be in the “XcoordxYcoord” (i.e., 10x10) format, 
#but you can also maintain your original spot names, for example, barcode names.
spatial_count
#### load the example spatial location data, 
load("./spa-data/spatial_location.RData")
spatial_location[1:4,]
#The spatial location data must be in the format of data frame while each row represents a spatial location,
# the first column represents the x coordinate and the second column represents the y coordinate.
# The rownames of the spatial location data frame should match exactly with the column names of the spatial_count.
spatial_location

2.细胞类型去卷积

1）Create an CARD object

library(CARD)

CARD_obj = createCARDObject(
    sc_count = sc_count,
    sc_meta = sc_meta,
    spatial_count = spatial_count,
    spatial_location = spatial_location,
    ct.varname = "cellType",
    ct.select = unique(sc_meta$cellType),
    sample.varname = "sampleInfo",
    minCountGene = 100,
    minCountSpot = 5) 
#The spatial data are stored in CARD_obj@spatial_countMat and CARD_obj@spatial_location 
#while the scRNA-seq data is stored in CARD_obj@sc_eset in the format of SingleCellExperiment.

2)Deconvolution using CARD

CARD_obj = CARD_deconvolution(CARD_object = CARD_obj)
#The results are stored in CARD_obj@Proportion_CARD.
CARD_obj@Proportion_CARD

3)Visualize the cell type proportion

colors = c("#FFD92F","#4DAF4A","#FCCDE5","#D9D9D9","#377EB8","#7FC97F","#BEAED4",
    "#FDC086","#FFFF99","#386CB0","#F0027F","#BF5B17","#666666","#1B9E77","#D95F02",
    "#7570B3","#E7298A","#66A61E","#E6AB02","#A6761D")
p1 <- CARD.visualize.pie(proportion = CARD_obj@Proportion_CARD,spatial_location = CARD_obj@spatial_location, colors = colors)
ggsave(p1,file="pie_plot.pdf",width=15,height=7)
#We do not recommend users to visualize this plot when the number of spots is > 500
pie_plot
## select the cell type that we are interested
ct.visualize = c("Acinar_cells","Cancer_clone_A","Cancer_clone_B","Ductal_terminal_ductal_like","Ductal_CRISP3_high-centroacinar_like","Ductal_MHC_Class_II","Ductal_APOL1_high-hypoxic","Fibroblasts")
## visualize the spatial distribution of the cell type proportion
p2 <- CARD.visualize.prop(
    proportion = CARD_obj@Proportion_CARD,        
    spatial_location = CARD_obj@spatial_location, 
    ct.visualize = ct.visualize,                 ### selected cell types to visualize
    colors = c("lightblue","lightyellow","red"), ### if not provide, we will use the default colors
    NumCols = 4)                                 ### number of columns in the figure panel
ggsave(p2,file="selected_cell_types.pdf",width=15,height=14)
selected_cell_types

4) Visualize the cell type proportion correlation

p3 <- CARD.visualize.Cor(CARD_obj@Proportion_CARD,colors = NULL)
ggsave(p3,file="correlation.pdf",width=14,height=14)
correlation

3.精细化空间图谱

1)Imputation on the newly grided spatial locations

CARD_obj = CARD.imputation(CARD_obj,NumGrids = 2000,ineibor = 10,exclude = NULL)
#The results are store in CARD_obj@refined_prop and CARD_obj@refined_expression

## Visualize the newly grided spatial locations to see if the shape is correctly detected. If not, the user can provide the row names of the excluded spatial location data into the CARD.imputation function
location_imputation = cbind.data.frame(x=as.numeric(sapply(strsplit(rownames(CARD_obj@refined_prop),split="x"),"[",1)),
    y=as.numeric(sapply(strsplit(rownames(CARD_obj@refined_prop),split="x"),"[",2)))
rownames(location_imputation) = rownames(CARD_obj@refined_prop)
library(ggplot2)
p4 <- ggplot(location_imputation, 
       aes(x = x, y = y)) + geom_point(shape=22,color = "#7dc7f5")+
theme(plot.margin = margin(0.1, 0.1, 0.1, 0.1, "cm"),
    legend.position="bottom",
    panel.background = element_blank(),
    plot.background = element_blank(),
    panel.border = element_rect(colour = "grey89", fill=NA, size=0.5))
ggsave(p4,file="newly_grid.pdf",width=14,height=14)
newly_grid

2)Visualize the cell type proportion at an enhanced resolution

p5 <- CARD.visualize.prop(
    proportion = CARD_obj@refined_prop,                         
    spatial_location = location_imputation,            
    ct.visualize = ct.visualize,                    
    colors = c("lightblue","lightyellow","red"),    
    NumCols = 4)    
ggsave(p5,file="enhanced_resolution_proportion.pdf",width=14,height=14)
enhanced_resolution_proportion

3)Visualize the marker gene expression at an enhanced resolution

p6 <- CARD.visualize.gene(
    spatial_expression = CARD_obj@refined_expression,
    spatial_location = location_imputation,
    gene.visualize = c("Tm4sf1","S100a4","Tff3","Apol1","Crisp3","CD248"),
    colors = NULL,
    NumCols = 6)
ggsave(p6,file="enhanced_resolution_gene_expression.pdf",width=14,height=7)
enhanced_resolution_gene_expression
#original gene expression
p7 <- CARD.visualize.gene(
    spatial_expression = CARD_obj@spatial_countMat,
    spatial_location = CARD_obj@spatial_location,
    gene.visualize = c("Tm4sf1","S100a4","Tff3","Apol1","Crisp3","CD248"),
    colors = NULL,
    NumCols = 6)
ggsave(p7,file="original_gene_expression.pdf",width=14,height=7)
original_gene_expression

4.不用单细胞作为参考对空转进行反卷积

不用单细胞数据，则需要一些能指明细胞类型的marker gene

wget https://github.com/YingMa0107/CARD/raw/master/data/markerList.RData

1)Create an CARDfree object

## load the marker gene list
load("./markerList.RData")
markerList

image.png
CARDfree_obj = createCARDfreeObject(
    markerList = markerList,
    spatial_count = spatial_count,
    spatial_location = spatial_location,
    minCountGene = 100,
    minCountSpot =5) 
#The spatial data are stored in CARDfree_obj@spatial_countMat and CARDfree_obj@spatial_location 
#while the marker list is stored in CARDfree_objj@markerList in the format of list.

2)Deconvolution using CARDfree

## deconvolution using CARDfree
CARDfree_obj = CARD_refFree(CARDfree_obj)
#The results are stored in CARDfree_obj@Proportion_CARD.

## One limitation of reference-free version of CARD is that the cell types inferred from CARDfree do not come with a cell type label. It might be difficult to interpret the results. 
print(CARDfree_obj@Proportion_CARD[1:2,])
CARDfree_obj@Proportion_CARD

3)Visualization of the results of CARDfree

colors = c("#FFD92F","#4DAF4A","#FCCDE5","#D9D9D9","#377EB8","#7FC97F","#BEAED4","#FDC086","#FFFF99","#386CB0","#F0027F","#BF5B17","#666666","#1B9E77","#D95F02","#7570B3","#E7298A","#66A61E","#E6AB02","#A6761D")
### In order to maximumply match with the original results of CARD, we order the colors to generally match with the results infered by CARD
CARDfree_obj@Proportion_CARD = CARDfree_obj@Proportion_CARD[,c(8,10,14,2,1,6,12,18,7,13,20,19,16,17,11,15,4,9,3,5)]
colnames(CARDfree_obj@Proportion_CARD) = paste0("CT",1:20)
p8 <- CARD.visualize.pie(CARDfree_obj@Proportion_CARD,CARDfree_obj@spatial_location,colors = colors)
ggsave(p8,file="pie_plot2.pdf",width=15,height=7)
pie_plot2

5.绘制单细胞分辨率的空间图谱

#### Note that here the shapeSpot is the user defined variable which indicates the capturing area of single cells. Details see above.
scMapping = CARD_SCMapping(CARD_obj,shapeSpot="Square",numCell=20,ncore=10)
print(scMapping)
scMapping
### spatial location info and expression count of the single cell resolution data
library(SingleCellExperiment)
MapCellCords = as.data.frame(colData(scMapping))
count_SC = assays(scMapping)$counts

library(ggplot2)
df = MapCellCords
colors = c("#8DD3C7","#CFECBB","#F4F4B9","#CFCCCF","#D1A7B9","#E9D3DE","#F4867C","#C0979F",
    "#D5CFD6","#86B1CD","#CEB28B","#EDBC63","#C59CC5","#C09CBF","#C2D567","#C9DAC3","#E1EBA0",
    "#FFED6F","#CDD796","#F8CDDE")
p9 = ggplot(df, aes(x = x, y = y, colour = CT)) + 
    geom_point(size = 3.0) +
    scale_colour_manual(values =  colors) +
    #facet_wrap(~Method,ncol = 2,nrow = 3) + 
        theme(plot.margin = margin(0.1, 0.1, 0.1, 0.1, "cm"),
              panel.background = element_rect(colour = "white", fill="white"),
              plot.background = element_rect(colour = "white", fill="white"),
    legend.position="bottom",
    panel.border = element_rect(colour = "grey89", fill=NA, size=0.5),
    axis.text =element_blank(),
    axis.ticks =element_blank(),
    axis.title =element_blank(),
    legend.title=element_text(size = 13,face="bold"),
    legend.text=element_text(size = 12),
    legend.key = element_rect(colour = "transparent", fill = "white"),
    legend.key.size = unit(0.45, 'cm'),
    strip.text = element_text(size = 15,face="bold"))+
                                guides(color=guide_legend(title="Cell Type"))
ggsave(p9,file="singlecell_map.pdf",width=15,height=7)
singlecell_map
最后编辑于 ：
©著作权归作者所有,转载或内容合作请联系作者
  • 人面猴
    序言：七十年代末，一起剥皮案震惊了整个滨河市，随后出现的几起案子，更是在滨河造成了极大的恐慌，老刑警刘岩，带你破解...
    沈念sama阅读 207,113评论 6赞 481
  • 死咒
    序言：滨河连续发生了三起死亡事件，死亡现场离奇诡异，居然都是意外死亡，警方通过查阅死者的电脑和手机，发现死者居然都...
    沈念sama阅读 88,644评论 2赞 381
  • 救了他两次的神仙让他今天三更去死
    文/潘晓璐 我一进店门，熙熙楼的掌柜王于贵愁眉苦脸地迎上来，“玉大人，你说我怎么就摊上这事。” “怎么了？”我有些...
    开封第一讲书人阅读 153,340评论 0赞 344
  • 道士缉凶录：失踪的卖姜人
    文/不坏的土叔 我叫张陵，是天一观的道长。 经常有香客问我，道长，这世上最难降的妖魔是什么？ 我笑而不...
    开封第一讲书人阅读 55,449评论 1赞 279
  • 港岛之恋（遗憾婚礼）
    正文 为了忘掉前任，我火速办了婚礼，结果婚礼上，老公的妹妹穿的比我还像新娘。我一直安慰自己，他们只是感情好，可当我...
    茶点故事阅读 64,445评论 5赞 374
  • 恶毒庶女顶嫁案：这布局不是一般人想出来的
    文/花漫 我一把揭开白布。 她就那样静静地躺着，像睡着了一般。 火红的嫁衣衬着肌肤如雪。 梳的纹丝不乱的头发上，一...
    开封第一讲书人阅读 49,166评论 1赞 284
  • 城市分裂传说
    那天，我揣着相机与录音，去河边找鬼。 笑死，一个胖子当着我的面吹牛，可吹牛的内容都是我干的。 我是一名探鬼主播，决...
    沈念sama阅读 38,442评论 3赞 401
  • 双鸳鸯连环套：你想象不到人心有多黑
    文/苍兰香墨 我猛地睁开眼，长吁一口气：“原来是场噩梦啊……” “哼！你这毒妇竟也来了？” 一声冷哼从身侧响起，我...
    开封第一讲书人阅读 37,105评论 0赞 261
  • 万荣杀人案实录
    序言：老挝万荣一对情侣失踪，失踪者是张志新（化名）和其女友刘颖，没想到半个月后，有当地人在树林里发现了一具尸体，经...
    沈念sama阅读 43,601评论 1赞 300
  • 护林员之死
    正文 独居荒郊野岭守林人离奇死亡，尸身上长有42处带血的脓包…… 初始之章·张勋 以下内容为张勋视角 年9月15日...
    茶点故事阅读 36,066评论 2赞 325
  • 白月光启示录
    正文 我和宋清朗相恋三年，在试婚纱的时候发现自己被绿了。 大学时的朋友给我发了我未婚夫和他白月光在一起吃饭的照片。...
    茶点故事阅读 38,161评论 1赞 334
  • 活死人
    序言：一个原本活蹦乱跳的男人离奇死亡，死状恐怖，灵堂内的尸体忽然破棺而出，到底是诈尸还是另有隐情，我是刑警宁泽，带...
    沈念sama阅读 33,792评论 4赞 323
  • 日本核电站爆炸内幕
    正文 年R本政府宣布，位于F岛的核电站，受9级特大地震影响，放射性物质发生泄漏。R本人自食恶果不足惜，却给世界环境...
    茶点故事阅读 39,351评论 3赞 307
  • 男人毒药：我在死后第九天来索命
    文/蒙蒙 一、第九天 我趴在偏房一处隐蔽的房顶上张望。 院中可真热闹，春花似锦、人声如沸。这庄子的主人今日做“春日...
    开封第一讲书人阅读 30,352评论 0赞 19
  • 一桩弑父案，背后竟有这般阴谋
    文/苍兰香墨 我抬头看了看天上的太阳。三九已至，却和暖如春，着一层夹袄步出监牢的瞬间，已是汗流浃背。 一阵脚步声响...
    开封第一讲书人阅读 31,584评论 1赞 261
  • 情欲美人皮
    我被黑心中介骗来泰国打工， 没想到刚下飞机就差点儿被人妖公主榨干…… 1. 我叫王不留，地道东北人。 一个月前我还...
    沈念sama阅读 45,618评论 2赞 355
  • 代替公主和亲
    正文 我出身青楼，却偏偏与公主长得像，于是被迫代替她去往敌国和亲。 传闻我的和亲对象是个残疾皇子，可洞房花烛夜当晚...
    茶点故事阅读 42,916评论 2赞 344