Console output from loop is printed inline with plot

[AlbertoFabbri93]

Description

Hi everybody,
I have the following code:

## Gene Set Enrichment Analysis

```{r}
# Set identity to tumor clusters
Idents(breast_cancer_patients_tumor_cells) <- tumor_cluster_name
# Set default assay to RNA
DefaultAssay(breast_cancer_patients_tumor_cells) <- "RNA"

# Run GSEA for each tumor cluster
for (cluster in levels(breast_cancer_patients_tumor_cells)) {
  tumor_cluster_number <- cluster
  print(paste("GSEA analysis of cluster:", tumor_cluster_number))

  # Get the markers for the tumor cluster
  tumor_cluster_markers <- tumor_markers_data %>%
    dplyr::filter(cluster == tumor_cluster_number)
  
  # Count the number of up-regulated genes
  num_upreg_genes <- tumor_cluster_markers %>%
    dplyr::filter(avg_log2FC > 0, p_val < 0.05) %>%
    nrow()
  # Count the number of down-regulated genes
  num_downreg_genes <- tumor_cluster_markers %>%
    dplyr::filter(avg_log2FC < 0, p_val < 0.05) %>%
    nrow()
  
  # How to create the gene list
  # http://yulab-smu.top/biomedical-knowledge-mining-book/faq.html#genelist
  # Gene list creation
  gsea_gene_list <- tumor_cluster_markers %>%
    dplyr::mutate(rank = rank(avg_log2FC, ties.method = "random")) %>%
    dplyr::arrange(desc(rank)) %>%
    with(setNames(avg_log2FC, gene))
  
  # Run GSEA
  gsea <- gseGO(geneList     = gsea_gene_list,
                    keyType      = "SYMBOL",
                    OrgDb        = org.Hs.eg.db, 
                    ont          = "BP",
                    minGSSize    = 10,
                    maxGSSize    = 500,
                    pvalueCutoff = 0.10,
                    verbose      = TRUE,
                    pAdjustMethod = "fdr")
  # Simplify GSEA results by grouping together or eliminating terms that are very similar or redundant
  gsea_simplified <- simplify(gsea, cutoff = 0.7)

  # Do not plot if no pathways have been found
  if (nrow(gsea_simplified@result) > 0) {
    # Prepare plot labels
    GSEAlabels <- labs(
      title = paste("Significant Pathways in Tumor Cluster", tumor_cluster_number), 
      subtitle = "by Gene Set Enrichment Analysis",
      caption = paste0("Based on ", num_upreg_genes," up-regulated genes and ", num_downreg_genes," down-regulated genes."))
    # Prepare plot aesthetics
    GSEAtheme <- theme(
      plot.title = element_text(hjust = 0.5, size = 20),
      plot.subtitle = element_text(hjust = 0.5, size = 15),
      axis.text.x = element_text(angle = 0, size = 11),
      axis.text.y = element_text(angle = 0, size = 8)) 
    # Make the GSEA plots
    GSEADotPlot <- dotplot(gsea_simplified, showCategory = 25) +
      GSEAlabels + labs(x = "Gene Ratio", y = "Biological Pathway") +
      GSEAtheme
    print(GSEADotPlot)
  } else {
    print(paste("No significant pathways found for cluster", tumor_cluster_number))
  }
}
```

Inside the for loop I am calling a function that analyze some data and print some information about the process to the console (verbose = TRUE). The problem is that the output from this function is only printed to the console for the first cluster, for the subsequent ones the console output is printed above the first plot.

Some images to show you what I mean

The console output with only the info about cluster 0:

The first plot has the verbose output for all the other runs of geoGO printed above it:

The other plots are normal. I tried to create a reproducible example but I am not sure on how to do that. Has anyone an idea on what is going on and how I can get all the verbose output to the console and not together with the first plot?
I tried to fix the problem with capture.output and sink but they are not able to capture the gseGO verbose output.

Thank you!

EDIT: The verbose output is printed above the first plot even when verbose = FALSE

[mcanouil]

You can try to make a code that output message(), warning(), and figures in a loop.

Without reproducible code, there is not much that can be done.

Note that this seems to be an issue with your settings in the context of knitr / R, rather than with Quarto CLI itself.

[mcanouil]

Indeed, I wanted to confirm that was the case here as well.

[AlbertoFabbri93]

In Quarto this is different that in R Markdown due to how options are set using YAML only. So here are the solution.

• A setup chunk in your .qmd has mentionned above with knitr::opts_chunk$set(message = NA) => that is the easiest way.

• You can do per chunk configuration, only valid for knitr, using

   #| message: !expr NA
  

I do not know if I am missing something here but I really can not get this to work. I have tried the first option by adding the code to a chunk above the loop and nothing changes. Same thing with adding the cell options to the top of the chunk with the loop:

[mcanouil]

Are you expecting this to work in the editor view?
Your screenshot is neither the rendered document, nor the terminal.

Try rendering your document independently of the IDE with quarto render your-file.qmd.

I believe your issue is unrelated to Quarto (or Rmarkdown).

[cderv]

RStudio Notebook view with inline output is not running knitr. So this a Rstudio issue on how they handle message and warning.

I missed that from your initial message - you should search rstudio IDE issues, and ask there. They probably do not support message: NA either or maybe it is something else.

Anything in notebook view is IDE handling the cells, not knitr.

[AlbertoFabbri93]

Ah ok, makes sense. Thank you very very much for all the explanations, I learned a lot!

[cderv]

You need to use the R/knitr way of setting the option in a code cell.

knitr::opts_chunk$set(message = NA)

NA is not a cross-engine valid value in Quarto.

See yihui.org/knitr/options.

With dev version of knitr, you can now set #| message: NA is a code cell.
yihui/knitr#2375

This will we converted from "NA" to NA by knitr and have the right effect. No need to change globally using the knit options.

I'll update YAML validation for this

[mcanouil]

That's awesome. (I can get rid of knitr::opts_chunk)

[cderv]

(I can get rid of knitr::opts_chunk)

Yes that this is the purpose 😄