從具有相等的nrow值和rownames的兩個scRNA-seq數據幀中刪除使用full_join生成的數據幀中的NA

Question

我一直在使用Log2數據框，如下所示：

library(dplyr)

str(df[1:10])

 $ 5W_Female_C#1_1    : num  0 0 0 0 0 ...
 $ 5W_Female_C#1_2    : num  2.28 0 0 0 0 ...
 $ 5W_Female_C#1_3    : num  0 0 0 0 0 ...
 $ 5W_Female_C#1_4    : num  2.15 0 0 1.79 0 ...
 $ 5W_Female_C#1_5    : num  0 0 0 0 0 ...
 $ 5W_Female_C#1_6    : num  0 0 0 0 0 ...
 $ 5W_Female_C#1_7    : num  0 0 0 1.41 0 ...
 $ 5W_Female_C#1_8    : num  0 0 0 0 0 ...
 $ 5W_Female_C#1_9    : num  0 0 0 0 0 ...
 $ 5W_Female_C#1_10   : num  0.18 0 0.18 0 0 ...

     dput(df[1:10,1:20])
structure(list(`5W_Female_C#1_1` = c(0, 0, 0, 0, 0, 0, 0, 1.23695175858808, 
2.08983709011962, 1.78366618090783), `5W_Female_C#1_2` = c(2.28362550060704, 
0, 0, 0, 0, 0.417920007811965, 0, 0, 4.23488447596799, 0), `5W_Female_C#1_3` = c(0, 
0, 0, 0, 0, 0, 0, 1.49722912878761, 2.95084163754915, 0), `5W_Female_C#1_4` = c(2.15088457130503, 
0, 0, 1.78993786898019, 0, 0.219091058246197, 0, 0, 3.48000655138599, 
0), `5W_Female_C#1_5` = c(0, 0, 0, 0, 0, 0, 0, 1.77610398807316, 
2.50182126542091, 0), `5W_Female_C#1_6` = c(0, 0, 0, 0, 0, 
0, 0, 3.01506932171765, 2.76107247078864, 1.42115596066222), 
    `5W_Female_C#1_7` = c(0, 0, 0, 1.40544784370754, 0, 0, 
    0, 1.12300395405482, 2.88009774972197, 0), `5W_Female_C#1_8` = c(0, 
    0, 0, 0, 0, 2.31875066934634, 0, 2.92257845650856, 3.34695688937888, 
    1.48284828306847), `5W_Female_C#1_9` = c(0, 0, 0, 0, 0, 
    0, 0, 1.61917821605907, 1.77273024776718, 2.09761079662642
    ), `5W_Female_C#1_10` = c(0.180147861158429, 0, 0.180147861158429, 
    0, 0, 0, 0, 0.180147861158429, 3.75103517666786, 0), `5W_Female_C#1_11` = c(0, 
    0, 0, 0.336854639125465, 0, 0, 0, 0, 2.7614980445501, 0), 
    `5W_Female_C#1_12` = c(0, 0, 0, 0, 0, 0, 0, 1.04404433270602, 
    3.39985467357243, 0), `5W_Female_C#1_13` = c(0, 0, 0, 0, 
    0, 0, 0, 0, 3.29484127140614, 1.12101540096137), `5W_Female_FGC#1_14` = c(0, 
    0, 0, 0, 0, 0, 0, 0, 3.00934717225925, 0), `5W_Female_C#1_15` = c(0.207892851641333, 
    0, 0, 0, 0, 1.26243315763135, 0, 0, 1.98294795515753, 0.829443681366591
    ), `5W_Female_C#1_16` = c(0, 0, 0, 1.15639661659767, 0, 
    0, 0, 1.8611613657534, 3.03509599789673, 0), `5W_Female_C#1_17` = c(0, 
    0, 0, 1.57385922157595, 0, 0, 0, 4.02874594222712, 0, 3.91513842592446
    ), `5W_Female_C#1_18` = c(0, 0, 0, 0, 0, 0, 0, 2.45180455572049, 
    3.72628628972067, 0), `5W_Female_C#1_19` = c(0.702214251010441, 
    0, 0, 0, 0, 0, 0, 1.41792000781196, 2.63853727921519, 1.17248751551013
    ), `5W_Female_C#1_20` = c(3.1243281350022, 0, 0, 0, 0, 
    0, 0, 0, 3.07963411236871, 0)), row.names = c("A1BG", "A1BG-AS1", 
"A1CF", "A2M", "A2M-AS1", "A2ML1", "A2MP1", "A4GALT", "AAAS", 
"AACS"), class = "data.frame")

小窗口的數據

這已根據基因/基因載體列表進行了子集化。

gene_list <- c("gene1","gene2","gene3","gene4","gene5")

隨后使用grep函數根據年齡進行后續子集化。 scdata4 <- as.data.frame(df[,grep("4W", colnames(df))]) scdata5 <- as.data.frame(df[,grep("5W", colnames(df))])

在此步驟之后，將行名稱（基因）放在名為gene的列下，使用：

tibble::rownames_to_column(df, var="gene")

最后，生成的兩個數據幀作為full_join函數的輸入給出，它們具有相同的nrow值rownames。

scdatajoin <- full_join(scdata4,scdata5, by = "gene")

當我使用as.matrix（）轉換為矩陣格式后，將此輸出插入到pheatmap函數中時，這里出現錯誤。

scdatajoin <- as.matrix(scdatajoin) pheatmap(scdatajoin, color=rev(brewer.pal(9,"RdBu")), main = "4plus5w")

我收到此錯誤：

Error in hclust(d, method = method) : NA/NaN/Inf in foreign function call (arg 11)

有人能告訴我如何糾正這個問題嗎？

Answer 1

在移除一個盡管存在於數據集中的基因之后，填充了NA而不是數值，我能夠運行該功能。 我通過使用包gplots的heatmap.2函數來實現這gplots ，在那里我可以看到這個基因測量值的缺失。