Error: make_cell_attr - log_umi contains NA, NaN, or infinite value in SCTransform

[unikill066]

Encountered an error while running SCTransform on a Xenium object created with LoadXenium. The error indicates that the log_umi cell attribute contains NA, NaN, or infinite values, which is preventing SCTransform from completing successfully.

Error:

Error in make_cell_attr(umi, cell_attr, latent_var, batch_var, latent_var_nonreg,  : 
  cell attribute "log_umi" contains NA, NaN, or infinite value

Code:

> library(Seurat)
> library(future)
> plan("multisession", workers = 10)
> library(ggplot2)
> library(tidyverse)
> library(arrow)
> library(dplyr)
> library(devtools)

> data.dir <- "/path/to/xenium/data/outs"

> xenium.obj <- LoadXenium(data.dir, fov = "fov", assay = "Xenium")
10X data contains more than one type and is being returned as a list containing matrices of each type.
Warning: Feature names cannot have underscores ('_'), replacing with dashes ('-')
Warning: Feature names cannot have underscores ('_'), replacing with dashes ('-')
Warning: Feature names cannot have underscores ('_'), replacing with dashes ('-')
Warning: Feature names cannot have underscores ('_'), replacing with dashes ('-')
Warning: Feature names cannot have underscores ('_'), replacing with dashes ('-')
Warning: Feature names cannot have underscores ('_'), replacing with dashes ('-')

> xenium.obj <- SCTransform(xenium.obj, assay = "Xenium", clip.range = c(-10, 10))
Calculating cell attributes from input UMI matrix: log_umi
Error in make_cell_attr(umi, cell_attr, latent_var, batch_var, latent_var_nonreg,  : 
  cell attribute "log_umi" contains NA, NaN, or infinite value

Please advise if there are specific pre-processing steps required for Xenium data or any known solutions to handle NA, NaN, or infinite values in log_umi.

[saketkc]

Do you have genes with zero counts in your data?

[HansLambda]

Did you remove empty cells as suggested in the vignette?

xenium.obj <- subset(xenium.obj, subset = nCount_Xenium > 0)