diffatac fails for hs1_chrR: object 'roiAnnodf' not found in DESeq2.Rmd #125

Summary

DESeq2.Rmd fails for the hs1_chrR genome with:

Error in merge.data.frame(resdf, roiAnnodf, by = "Geneid") :
  object 'roiAnnodf' not found

This crashes the diffatac rule for every peakcaller/method/contrast combination when genome: hs1_chrR (also affects hs1).

Root cause

The pieplot chunk in workflow/scripts/DESeq2.Rmd only builds roiAnnodf when params$genome is "hg38" or "mm10":

if (params$genome == "hg38") { ... roiAnnodf <- as.data.frame(roiAnno) ... }
if (params$genome == "mm10") { ... roiAnnodf <- as.data.frame(roiAnno) ... }

There is no branch for hs1/hs1_chrR, so roiAnnodf is never created, and the later result chunk crashes at:

resdf_w_anno <- merge(resdf, roiAnnodf, by = "Geneid")

This is the same class of gap fixed in #123 for ccbr_annotate_bed.R/ccbr_annotate_peaks.R (T2T CHM13v2.0 TxDb support), but it was missed in DESeq2.Rmd when hs1_chrR support was added in #120.

Proposed fix

Add an hs1/hs1_chrR branch to the pieplot chunk in DESeq2.Rmd, reusing the T2T TxDb (/opt2/annotation/TxDb.Hsapiens.NCBI.T2T.CHM13v2.0.sqlite) and org.Hs.eg.db pattern already used in ccbr_annotate_bed.R/ccbr_annotate_peaks.R, including filling ENSEMBL/SYMBOL/GENENAME with NA when ChIPseeker can't map T2T gene IDs (since SYMBOL is referenced later in the volcano plot chunk).

Affected

• workflow/scripts/DESeq2.Rmd
• Snakemake diffatac rule (all peakcaller × method × contrast combinations for hs1_chrR/hs1 genomes)

_{⚡ Generated using AI ⚡}