Error music_basis and music_prop: Error in h(simpleError(msg, call)) : error in evaluating the argument 'i' in selecting a method for function '[': 'x' must be an array of at least two dimensions

[YColino]

Hi,

I am having the following error when I run both music_basis and music_prop functions:

Error in h(simpleError(msg, call)) : 
  error in evaluating the argument 'i' in selecting a method for function '[': 'x' must be an array of at least two dimensions

If I try with the example data it works but when I use my data it does not.

Here is my code:

sc.basis <- music_basis( x=sce, non.zero = TRUE,  markers = NULL,   clusters='cellType',samples='sampleID', select.ct = NULL )

Est.prop <- music_prop(bulk.mtx = bulk.m, sc.sce = sce, clusters = 'cellType',  samples = 'sampleID',   select.ct = c("Epithelial cells",  "Fibroblasts" ,  "Macrophages" ,"T cells"   , "Endothelial cells" ,"Stroma cells" , "Neutrophils" ), markers= NULL)

The sce looks like this:

 >sce
class: SingleCellExperiment 
dim: 32845 7958 
metadata(0):
assays(1): counts
rownames(32845): Xkr4 Gm1992 ... Trbj2.5 Vamp9
rowData names(0):
colnames(7958): 678478_1 791068_1 ... 820364_2 458736_2
colData names(2): sampleID cellType
reducedDimNames(0):
mainExpName: NULL
altExpNames(0):

The bulk.m:

>head(bulk.m)
              noise_expression_fresh noise_expression_24h
0610009B22Rik           1.268665e-04         1.807174e-04
0610010F05Rik           1.014932e-04         1.161755e-04
0610030E20Rik           6.343326e-05         1.549007e-04
0610037L13Rik           6.343326e-05         9.035872e-05
0610040B10Rik           0.000000e+00         1.290839e-05
0610040J01Rik           1.649265e-04         1.290839e-04
> class(bulk.m)
[1] "matrix" "array" 
>dim( bulk.m)
[1] 10136     2
> str( bulk.m)
 num [1:10136, 1:2] 1.27e-04 1.01e-04 6.34e-05 6.34e-05 0.00 ...
 - attr(*, "dimnames")=List of 2
  ..$ : chr [1:10136] "0610009B22Rik" "0610010F05Rik" "0610030E20Rik" "0610037L13Rik" ...
  ..$ : chr [1:2] "noise_expression_fresh" "noise_expression_24h"

I think the problem is in the single cell cause when I use the example bulk RNA provided I get the same problem, but if I use my bulk but the example sc data it says "Too few common genes!".

Any guidance on how to solve the error would be much appreciated!

Thanks, Yolanda

[xoelmb]

Hi! I've been dealing with a similar problem for a couple of days, so I hope the way I went around it works for you.

I got the same error in any music_basis/music_prop/music2_prop as you did when the SCE contained only one value in the samples variable. In my case, I was testing a 1-donor version of a 4-donor dataset when this showed up. So check the data in the sce column that corresponds to your samples parameter in the function call. I went through the code in music_basis and I think it makes sense because the creation of the S matrix expects somehow both samples and celltypes to compute the relative abundances. Anyhow, is MuSiC intended for 1 sample datasets?

So, I went for the full dataset and the "Too few common genes!" problem also appeared. I checked: my data have some 30000 genes present in both bulk and single-cell datasets. I think I solved partially the problem with this piece of code, which basically, subsets and resorts the data to common genes:

c.genes <- sort(intersect(rownames(control.mtx), rownames(sce)))

sce <- sce[c.genes,]

control.mtx <- as.matrix(control.mtx[c.genes,])
# case.mtx <- as.matrix(case.mtx[c.genes,])

dim(sce)
dim(control.mtx)
# dim(case.mtx)

Now, the Too few common genes! error is gone, though I think this could/should be fixed in the package itself, no?

PS: music2_prop is still complaining about non-conformable arrays, but I don't think it's related to these problems. Error in MOD0 + mod0: non-conformable arrays