Error when running ```restrictCorpus```

[cathalgking]

I encounter the error below when running restrictCorpus on a 10x Visium dataset (TMA) contained within a SpatialFeatureExperiment object.
Can you advise how to get past this?

vis1 <- restrictCorpus(cd, removeAbove=1.0, removeBelow = 0.05, nTopOD = 1000)

Removing 0 genes present in 100% or more of pixels...
18085 genes remaining...
Removing 5615 genes present in 5% or less of pixels...
12470 genes remaining...
Restricting to overdispersed genes with alpha = 0.05...
Converting to sparse matrix ...
Calculating variance fit ...
Using lm ...
Error in lm.fit(x, y, offset = offset, singular.ok = singular.ok, ...) : 
  0 (non-NA) cases

data

vis1
class: SpatialFeatureExperiment 
dim: 18085 5210 
metadata(0):
assays(1): counts
rownames(18085): SAMD11 NOC2L ... MT-ND6 MT-CYB
rowData names(8): symbol Feature.Type ... Median.Normalized.Average.Counts_sample01 Barcodes.Detected.per.Feature_sample01
colnames(5210): AACAATCCGAGTGGAC-1 AACACCAGCCTACTCG-1 ... TGTTGGTGCGGAATCA-1 TGTTGGTGGACTCAGG-1
colData names(4): in_tissue array_row array_col sample_id
reducedDimNames(0):
mainExpName: NULL
altExpNames(0):
spatialCoords names(2) : pxl_col_in_fullres pxl_row_in_fullres
imgData names(4): sample_id image_id data scaleFactor

unit: full_res_image_pixel
Geometries:
colGeometries: spotPoly (POLYGON) 

Graphs:
sample01: col: visium

[cathalgking]

I read the data into a SpatialExperiment instead and did not get the error at restrictCorpus

vis1 <- restrictCorpus(cd, removeAbove=1.0, removeBelow = 0.05, nTopOD = 1000)
Removing 0 genes present in 100% or more of pixels...
18085 genes remaining...
Removing 5615 genes present in 5% or less of pixels...
12470 genes remaining...
Restricting to overdispersed genes with alpha = 0.05...
Calculating variance fit ...
Using gam with k=5...
2262 overdispersed genes ... 
 Using top 1000 overdispersed genes.

However, now get a similar error at fitLDA

ldas <- fitLDA(t(as.matrix(vis1)), Ks = c(2))
Error in fitLDA(t(as.matrix(vis1)), Ks = c(2)) : 
  Each row (pixel) of `counts` needs to contain at least one non-zero entry

I tried to remove empty rows but still get the same error.

> vis1_filtered <- vis1[rowSums(vis1) > 0, ]
> 
> ldas <- fitLDA(t(as.matrix(vis1_filtered)), Ks = c(2))
Error in fitLDA(t(as.matrix(vis1_filtered)), Ks = c(2)) : 
  Each row (pixel) of `counts` needs to contain at least one non-zero entry

[JEFworks]

You are on the right track. Do note the error message that “Each row (pixel) of `counts` needs to contain at least one non-zero entry” `fitLDA` accepts a transposed version ie. `t()` of `vis1` so the row of the matrix passed into `fitLDA` is the column of the original `vis1` matrix. You therefore need to remove the empty columns, which correspond to pixels or spots in your Visium data. Hope that helps. Jean Fan, PhD Associate Professor in Biomedical Engineering Center for Computational Biology Johns Hopkins University JEFworks Lab: jef.works<https://jef.works/> On Aug 27, 2025, at 8:00 PM, Cathal King ***@***.***> wrote: External Email - Use Caution [https://avatars.githubusercontent.com/u/32261323?s=20&v=4]cathalgking left a comment (JEFworks-Lab/STdeconvolve#72)<#72 (comment)> I read the data into a SpatialExperiment instead and did not get the error at restrictCorpus vis1 <- restrictCorpus(cd, removeAbove=1.0, removeBelow = 0.05, nTopOD = 1000) Removing 0 genes present in 100% or more of pixels... 18085 genes remaining... Removing 5615 genes present in 5% or less of pixels... 12470 genes remaining... Restricting to overdispersed genes with alpha = 0.05... Calculating variance fit ... Using gam with k=5... 2262 overdispersed genes ... Using top 1000 overdispersed genes. However, now get a similar error at fitLDA ldas <- fitLDA(t(as.matrix(vis1)), Ks = c(2)) Error in fitLDA(t(as.matrix(vis1)), Ks = c(2)) : Each row (pixel) of `counts` needs to contain at least one non-zero entry I tried to remove empty rows but still get the same error.

vis1_filtered <- vis1[rowSums(vis1) > 0, ] ldas <- fitLDA(t(as.matrix(vis1_filtered)), Ks = c(2))

Error in fitLDA(t(as.matrix(vis1_filtered)), Ks = c(2)) : Each row (pixel) of `counts` needs to contain at least one non-zero entry — Reply to this email directly, view it on GitHub<#72 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ACEPMSGNA4JWENJMXWSI2WD3PZBBBAVCNFSM6AAAAACE5VTJD2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTEMZQGIYDOMJXGE>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>