Error with getMarkerFeatures(): Error in colMeans(m) : 'x' must be numeric

[zrcjessica]

Hello,

I have snATAC-seq from multiple samples, and have already labeled the cell types and called peaks on the dataset. Within each cell type, I want to find differential marker peaks between groups A and B (high addiction vs. low addiction).

My command is as follows:

# subset for cells belonging to specific cell type
idxSample <- BiocGenerics::which(proj2$Clusters2 %in% names(table(proj2$Clusters2))[1])
cellsSample <- proj2$cellNames[idxSample]
proj2[cellsSample, ]

markersPeaks <- getMarkerFeatures(
    ArchRProj = proj2[cellsSample, ], 
    useMatrix = "PeakMatrix", 
    groupBy = "Addiction",
    useGroups = "high",
    bgdGroups = "low",
  bias = c("TSSEnrichment", "log10(nFrags)", "Library", "Sample"),
  testMethod = "wilcoxon"
)

I get the following error:

[ArchR-getMarkerFeatures-6091249a108f-Date-2022-01-12_Time-16-15-01.log](https://github.com/GreenleafLab/ArchR/files/7859040/ArchR-getMarkerFeatures-6091249a108f-Date-2022-01-12_Time-16-15-01.log)

ArchR logging to : ArchRLogs/ArchR-getMarkerFeatures-6091249a108f-Date-2022-01-12_Time-16-15-01.log
If there is an issue, please report to github with logFile!
2022-01-12 16:15:02 : Matching Known Biases, 0.015 mins elapsed.
Error in colMeans(m) : 'x' must be numeric
Calls: getMarkerFeatures ... SimpleList -> lapply -> FUN -> .summarizeColStats -> colMeans
In addition: Warning messages:
1: package ‘nabor’ was built under R version 4.1.1 
2: In mean.default(sort(x, partial = half + 0L:1L)[half + 0L:1L]) :
  argument is not numeric or logical: returning NA
3: In mean.default(sort(x, partial = half + 0L:1L)[half + 0L:1L]) :
  argument is not numeric or logical: returning NA
4: In mean.default(sort(x, partial = half + 0L:1L)[half + 0L:1L]) :
  argument is not numeric or logical: returning NA
5: In mean.default(sort(x, partial = half + 0L:1L)[half + 0L:1L]) :
  argument is not numeric or logical: returning NA
Execution halted
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100    82    0     2  100    80      5    219 --:--:-- --:--:-- --:--:--   225

I ran colMeans() on my PeakMatrix to make sure everything was looking okay:

peakMatrix <- getMatrixFromProject(proj, "PeakMatrix")
colMeans(assay(peakMatrix))

After wading through the source code, I realized that the problem was that I had passed qualitative variables to the bias argument which causes the matrix m in colMeans(m) a character matrix rather than a numeric matrix. I tested the source code within Jupyter notebook on my data and it worked just fine. I resubmitted my job removing the qualitative variables and including only the default bias variables (bias = c("TSSEnrichment", "log10(nFrags)")). However, I got the same error again. Now I'm not sure why, especially since the same data worked just fine within Jupyter notebook.

ArchR log file attached. I'm using ArchR version 1.0.1, which is consistent with the version of the source code I was working with.

Help is much appreciated!
ArchR-getMarkerFeatures-6091249a108f-Date-2022-01-12_Time-16-15-01.log