Error in function heatmap: 'x' must be a numeric matrix

Question

[f.morgana]

Hi all,

I have been struggling with the heatmap function for which I continuously receive the error: 'x' must be a numeric matrix. I have been trying to perform section 4 ("Identifying temporally expressed genes") of the following slingshot vignette: http://www.bioconductor.org/packages/devel/bioc/vignettes/slingshot/inst/doc/vignette.html

I have executed the following lines of code, where "thymus_sce" is my SCE object:

library(tradeSeq)

# fit negative binomial GAM
sim <- fitGAM(thymus_sce)

# test for dynamic expression
ATres <- associationTest(thymus_sce)

topgenes <- rownames(ATres[order(ATres$pvalue), ])[1:250]
pst.ord <- order(thymus_sce$slingPseudotime_1, na.last = NA)
heatdata <- assays(thymus_sce)$counts[topgenes, pst.ord]
heatclus <- thymus_sce$GMM[pst.ord]

All of the above I got to work, until I get to:

heatmap(log1p(heatdata), Colv = NA,
        ColSideColors = brewer.pal(9,"Set1")[heatclus])

Here, I get:

Error in heatmap(log1p(heatdata), Colv = NA, ColSideColors = brewer.pal(9, : 'x' must be a numeric matrix

I checked the class of heatdata:

class(heatdata)

which gives me

[1] "dgCMatrix"
attr(,"package")
[1] "Matrix"

So I suppose it is a matrix, right? At least, a sparse one... Is that the problem? What am I doing wrong and how could I get this to work?

Your help would be much appreciated, if any more information is needed, please let me know!

Kind regards, Florencia

[JC]

some R variables are not a matrix, could be a list, dataset or object, try to enforce matrix format with:

heatmap(as.matrix(log1p(heatdata)), Colv = NA,
    ColSideColors = brewer.pal(9,"Set1")[heatclus])

[f.morgana]

Hi JC,

Many thanks for your rapid response! I tried your suggestion, it seems to fix my previous error but now I get another one stating:

Error in heatmap(as.matrix(log1p(heatdata)), Colv = NA, ColSideColors = brewer.pal(9,  : 
  'ColSideColors' must be a character vector of length ncol(x)

To be honest, I just copy/pasted the lines of code from the vignette, so not entirely sure how to modify the ColSideColors argument... Any idea?

[Friederike]

you can either remove the entire argument (remove ColSideColors = ....) or you can make sure that heatclus is the same value as ncol(log1p(heatdata))

[JC]

that part is defining the colors, so you need a vector with one value per column, you can change it as:

ColSideColors = brewer.pal(9,"Set1")(ncol(as.matrix(heatdata)))