IQR计算和表格

Question

假设有一个数据框如下所示：

>                         A           B              C             D             E
>        TGFBI      0.027442647  9.756301e-03  0.0056374607  0.0248263371  0.0056703467
>        OLFM4      0.022665292 -1.906351e-03 -0.0135277027  0.0551336843  0.0001602728
>        CD177      0.029256398  2.259310e-03 -0.0218761784  0.0008816893 -0.0138302621
>        LCN2       0.024944813  1.838820e-02 -0.0058928266  0.0440654781 -0.0108800098
>        CEACAM8    0.029996651  3.432132e-02 -0.0251011180  0.0370074902 -0.0138822167
>        HLA-DPB1   0.028016101  3.483277e-02 -0.0081639565  0.0223873901  0.0103236673
>        DEFA3     -0.031190483  4.124520e-02 -0.0410158867  0.0607274629  0.0158699504

我想逐行计算IQR和之前计算的IQR (>=，<=)之外的abs(值)的数量，并将其输出到最终的表中。

换句话说，我想计算每一行中有多少个极值。

data.frame包含174列和8000行。

Answer 1

正如乔戈在他的评论中指出的那样，要求第一和第三个四分位数以外的分数没有太大意义。如果您将'outlier‘定义为四分位数外的点，则可以对以下代码进行一些修改：

# sample data
df <- read.table(text = "          A           B              C             D             E
TGFBI      0.027442647  9.756301e-03  0.0056374607  0.0248263371  0.0056703467
           OLFM4      0.022665292 -1.906351e-03 -0.0135277027  0.0551336843  0.0001602728
           CD177      0.029256398  2.259310e-03 -0.0218761784  0.0008816893 -0.0138302621
           LCN2       0.024944813  1.838820e-02 -0.0058928266  0.0440654781 -0.0108800098
           CEACAM8    0.029996651  3.432132e-02 -0.0251011180  0.0370074902 -0.0138822167
           HLA-DPB1   0.028016101  3.483277e-02 -0.0081639565  0.0223873901  0.0103236673
           DEFA3     -0.031190483  4.124520e-02 -0.0410158867  0.0607274629  0.0158699504", 
           header = TRUE, stringsAsFactors = FALSE)

# apply a custom function to rows of the data frame
apply(df, 1, function(x){
    qrt <- quantile(x, c(0.25, 0.75))
    iqr <- qrt[2] - qrt[1]
    out1 <- sum(x < qrt[1] - 1.5*iqr) # or use other value than 1.5*iqr
    out2 <- sum(x > qrt[2] + 1.5*iqr)
    return(out1 + out2) # returns just the number of outliers
})

#TGFBI    OLFM4    CD177    LCN2     CEACAM8  HLA-DPB1 DEFA3 
#0        0        1        0        0        0        0