R矩阵的元素转换为字符 - 如何挽救唯一值？

I'm working with a large matrix, M, which contains sample by gene data. Certain elements contain multiple concatenated entries, but the vectors have been converted to characters:

我正在使用一个大矩阵M，它包含基因数据样本。某些元素包含多个连接条目，但向量已转换为字符：

        geneA    gene2             
sample1    NA    NA                
sample2    "c(\"siteX\", \"siteY\")"    "0"
sample3    "siteZ"    "0"

So when I call unique(M[,'geneA']) I get:

因此，当我调用unique（M [，'geneA']）时，我得到：

NA    "c(\"siteX\", \"siteY\")"    "siteZ"

Is there any way to 'de-characterize' the matrix so that I can obtain all of the unique values for geneA when I run my code? Or would it be better to focus instead on extracting and manipulating the elements of interest using regular expressions?

有没有办法'去表征'矩阵，以便在运行代码时可以获得geneA的所有唯一值？或者更好地关注使用正则表达式提取和操纵感兴趣的元素？

Thanks in advance!

提前致谢！

edit:

编辑：

> dput(tmp)
structure(c(NA, "c(\"siteX\", \"siteY\")", 
"siteZ", NA, "0", "0"), .Dim = c(3L, 2L), .Dimnames = list(
c("sample1", "sample2", "sample3"), c("geneA", 
"gene2")))

1 个解决方案

#1

If you just want the unique geneA values:

如果您只想要独特的geneA值：

df = structure(c(NA, "c(\"siteX\", \"siteY\")", "siteZ", NA, "0", "0"), .Dim = c(3L, 2L), 
               .Dimnames = list(c("sample1", "sample2", "sample3"), c("geneA", "gene2")))
df = data.frame(df,stringsAsFactors=F)
df$geneA = as.character(df$geneA)
geneA = unlist(sapply(df$geneA, 
                      function(x) 
                      { sapply(strsplit(x,",")[[1]], 
                               function(x) { sub(".*\"(\\w+)\".*", "\\1", x,perl=TRUE) }      )  }))
names(geneA) = NULL
unique(geneA)

Nicer ordering and cleaning into a data frame with removal of NA geneA's:

通过删除NA geneA来更好地排序和清理数据框：

df = structure(c(NA, "c(\"siteX\", \"siteY\")", "siteZ", NA, "g1", "g2"), .Dim = c(3L, 2L), 
               .Dimnames = list(c("sample1", "sample2", "sample3"), c("geneA", "gene2")))
df = data.frame(df,stringsAsFactors=F)
df$geneA = as.character(df$geneA)
require(plyr)
ddply(df, "geneA", function(x)
    { 
      if(!is.na(x))
      {
        geneAs = sapply(strsplit(x$geneA,",")[[1]], function(y) { sub(".*\"(\\w+)\".*", "\\1", y,perl=TRUE) } );
        return(data.frame("geneA"= geneAs, "gene2" = rep(x$gene2[1],length(geneAs)) )) 
      } else return(NULL)
    } )

#1