从相关矩阵中寻找具有良好相关性的基因

Question

我有矩阵文件，基本上是跨不同细胞类型的基因之间的spearman相关矩阵。所以现在我试图找出哪一组基因或一组相关值大于0.6的基因，如果我设定这个阈值的话。我怎么能这么做？我在张贴我的数据子集。这是一个502 x 502矩阵。

        ACTL6B   ACTR5   ACTR6
ACTL6B  1        0.6        -0.4
ACTR5   0.4        1        -0.3
ACTR6  -0.4      -0.3         1

所以我不想要同一组基因之间的相关性，那就是1，我想要另一种比较。比方说，ACTL6B和ACTR5的相关性是0.6。我想保留这些价值观和基因。

Answer 1

下面是一个示例：

mat <- cor(longley)  # example 7 x 7 correlation matrix

# Find indices of correlations greater than 0.6
idx <- which(mat > 0.6 & lower.tri(mat), arr.ind = TRUE)

# names of the resulting variables
cbind(rownames(idx), colnames(mat)[idx[, 2]])

由于lower.tri的存在，对角和上矩阵的所有值都被忽略了。

结果：

      [,1]         [,2]          
 [1,] "GNP"        "GNP.deflator"
 [2,] "Unemployed" "GNP.deflator"
 [3,] "Population" "GNP.deflator"
 [4,] "Year"       "GNP.deflator"
 [5,] "Employed"   "GNP.deflator"
 [6,] "Unemployed" "GNP"         
 [7,] "Population" "GNP"         
 [8,] "Year"       "GNP"         
 [9,] "Employed"   "GNP"         
[10,] "Population" "Unemployed"  
[11,] "Year"       "Unemployed"  
[12,] "Year"       "Population"  
[13,] "Employed"   "Population"  
[14,] "Employed"   "Year"