[BioC] Extracting rows where any column has a non-zero value

[Sean Davis]

On Fri, Dec 5, 2008 at 12:28 AM, Nathan S. Watson-Haigh
<nhaigh at sheffield.ac.uk> wrote:
> I have what I think should be really obvious, but seem to keep banging my
> head against a wall....ouch!
>
> I want generate a heatplot for a set of genes from a contrast used in my
> limma analysis. However, I want to limit the heatplot to only those arrays
> used in the contrast. So I thought I'd be able to use the contrast and
> design matrix to do just that. However, I'm having problems.....
>
>
> My design matrix:
>> design
>   H.N.N L.N.N H.Y.N L.Y.N H.Y.Y L.Y.Y
> 1      1     0     0     0     0     0
> 2      1     0     0     0     0     0
> 3      1     0     0     0     0     0
> 4      0     1     0     0     0     0
> 5      0     1     0     0     0     0
> 6      0     1     0     0     0     0
> 7      0     0     1     0     0     0
> 8      0     0     1     0     0     0
> 9      0     0     1     0     0     0
> 10     0     0     0     1     0     0
> 11     0     0     0     1     0     0
> 12     0     0     0     1     0     0
> 13     0     0     0     0     1     0
> 14     0     0     0     0     1     0
> 15     0     0     0     0     1     0
> 16     0     0     0     0     0     1
> 17     0     0     0     0     0     1
> 18     0     0     0     0     0     1
> attr(,"assign")
> [1] 1 1 1 1 1 1
> attr(,"contrasts")
> attr(,"contrasts")$TG
> [1] "contr.treatment"
>
>
> My contrast matrix:
>> cont.matrix
>       Contrasts
> Levels  Helin_Acclimation Helin_Freezing Leitrim_Acclimation
> Leitrim_Freezing
>  H.N.N                -1              0                   0
> 0
>  L.N.N                 0              0                  -1
> 0
>  H.Y.N                 1             -1                   0
> 0
>  L.Y.N                 0              0                   1
> -1
>  H.Y.Y                 0              1                   0
> 0
>  L.Y.Y                 0              0                   0
> 1
>       Contrasts
> Levels  Acclimation Freezing Diff_Acclimation Diff_Freezing Population
>  H.N.N          -1        0                1             0         -1
>  L.N.N          -1        0               -1             0          1
>  H.Y.N           1       -1               -1             1         -1
>  L.Y.N           1       -1                1            -1          1
>  H.Y.Y           0        1                0            -1         -1
>  L.Y.Y           0        1                0             1          1
>
>
> So I want to do this in a loop for all the contrasts I made:
> for (mycoef in colnames(cont.matrix)) {
>  pval.max <- 0.05
>  maxheatplot.row <- 100
>
>  # get stats for this coef
>  TT <- topTable(fit2, adjust.method="BH", coef=mycoef, n=nrow(fit2))
>
>  # Just use those with an adj.p-value of < pval.max
>  TT <- TT[which(TT$adj.P.Val < pval.max),]
>
>  # make a heatplot for for the selected genes and include only those arrays
> used in the contrast
>  if (nrow(TT) > 1) {
>  heatplot.rows <- if(nrow(TT) > maxheatplot.row) { maxheatplot.row } else {
> nrow(TT) }
>
>  # index of those genes to include in the heatplot
>  index <- as.numeric(row.names(TT[c(1:heatplot.rows),]))
>
>  # Use the Contrast and design matrix to identify the arrays used in the
> contast
>  arrays <- which(design[,labels(which(cont.matrix[,mycoef] != 0))] !=0)
>
>  png(file = paste("HP_", gsub(":", "_", mycoef), ".png", sep=""), type =
> "cairo1", width=10, height=10, units="in", res=300)
>  heatplot(eSet[index,arrays], labCol=sampleNames(eSet)[arrays],
> labRow=TT$Symbol, margins=c(10,10), dend="row")
>  dev.off()
>  }
> }
>
>
> In the section dealing with creating the heatplot, the following code pulls
> out a subset of the design matrix:
>> design[,labels(which(cont.matrix[,mycoef] != 0))]
>   H.N.N H.Y.N
> 1      1     0
> 2      1     0
> 3      1     0
> 4      0     0
> 5      0     0
> 6      0     0
> 7      0     1
> 8      0     1
> 9      0     1
> 10     0     0
> 11     0     0
> 12     0     0
> 13     0     0
> 14     0     0
> 15     0     0
> 16     0     0
> 17     0     0
> 18     0     0
>
> I now want to pull out the row names which have a non-zero value in any one
> of the columns......can anyone help??

> a <- matrix(c(1,0,0,0,0,1),nc=2)
> a
     [,1] [,2]
[1,]    1    0
[2,]    0    0
[3,]    0    1
> which(rowSums(a)>0)
[1] 1 3

Does that do approximately what you want?

Sean