Seurat::DoHeatmap 的謎之報錯

最近在學(xué)習(xí)Seurat教程（Seurat版本3.0.2），發(fā)現(xiàn)運行到倒數(shù)第二步做熱圖展示marker genes時總是報錯

主要問題：

DoHeatmap(pbmc, features = top10$gene) + NoLegend()

運行結(jié)果：

Warning message:
In DoHeatmap(pbmc, features = top10$gene) :
The following features were omitted as they were not found in the scale.data slot for the RNA assay: NEAT1, SLC25A39, AC090498.1, CTSS, DNAJB1, HSP90AA1, PRF1, NKG7, KLRD1, CD8B, DUSP4, CD8A, CCL5, CD27, GZMK, CD69, TNFAIP3, ZFP36L2, IL7R

圖片倒是生成了，但是數(shù)了一下基因數(shù)，確實少了很多，尤其是Cluster 0 和1,都沒幾個剩下的了

1565194413212.png

嘗試自己調(diào)，一晚上也就調(diào)到這個程度：

1565194545049.png

好丑，還是繼續(xù)研究下包里的代碼吧/(ㄒoㄒ)/~~

DoHeatMap

看了一下代碼，里面有這樣一句：

  if (any(!features %in% possible.features)) {
    bad.features <- features[!features %in% possible.features]
    features <- features[features %in% possible.features]
    if (length(x = features) == 0) {
      stop("No requested features found in the ", 
           slot, " slot for the ", assay, " assay.")
    }
    warning("The following features were omitted as they were not found in the ", 
            slot, " slot for the ", assay, " assay: ", 
            paste(bad.features, collapse = ", "))

這個bad.features來自features和possible.feature的比對，之后判斷若features內(nèi)元素個數(shù)不為零則warning："The following features were omitted as they were not found in the ······“

function (object, features = NULL, cells = NULL, group.by = "ident", 
          group.bar = TRUE, disp.min = -2.5, disp.max = NULL, slot = "scale.data", 
          assay = NULL, label = TRUE, size = 5.5, hjust = 0, angle = 45, 
          raster = TRUE, draw.lines = TRUE, lines.width = NULL, group.bar.height = 0.02, 
          combine = TRUE) 

往回找，原來DoHeatMap（）有個默認(rèn)參數(shù)slot = "scale.data"，自動用scale.data的數(shù)據(jù)畫圖。
看一下這個Seurat對象的結(jié)構(gòu)：

1565195235734.png

哈哈，發(fā)現(xiàn)問題了，原來counts （應(yīng)該是raw read count）和data 都是20647行，每行對應(yīng)一個基因，但scale.data只有2000行。

> pbmc <- ScaleData(
+   object = pbmc,
+   do.scale = TRUE,
+   do.center = FALSE,
+   vars.to.regress = c("percent.mt"))

重新ScaleData()一下，發(fā)現(xiàn)數(shù)據(jù)沒有任何改變····

算了，試試用data作圖：

DoHeatmap(pbmc, features = top10$gene，slot = "data") 

1565195494082.png

還是很難看，間接證明Seurat包開發(fā)者的可視化功力非同一般。但是沒有報warning，而且可以看到消失的基因都出現(xiàn)了。

試試直接給scale.data賦值：

pbmc@assays$RNA@scale.data <- scale(pbmc@assays$RNA@data, scale = TRUE)
DoHeatmap(pbmc, features = top10$gene,size = 0.5,slot = "scale.data") + NoLegend()

這次也沒有報warning，配色還是有點難看，大概是scale參數(shù)設(shè)置的不同，想完全復(fù)現(xiàn)可能還得查ScaleData的代碼了。不過基本上能用了。增加了height，避免文字?jǐn)D在一起，其實最好是保存成ggplot的對象，拼圖的時候再改。

1565195816712.png

總結(jié)：

產(chǎn)生這個問題的主要原因是ScaleData()指令生成scale.data（封裝在Seurat對象里）中基因數(shù)減少。

因為Seurat對象封裝了好幾層，并且對S4對象的操作也不太熟悉，所以一開始不太容易發(fā)現(xiàn)原因。

另外，操作多個Seurat對象時注意偶爾rm（）+gc（）清空內(nèi)存，我24G內(nèi)存都好幾次99%內(nèi)存占用(╯‵□′)╯︵┻━┻