RNA-seq¶

RNA runs default to HISAT2 mapping, keep duplicate reads, use MAPQ 15 filtering, and create separate plus- and minus-strand BigWigs.

omnomnomics rna \
  -i EXPERIMENT \
  -g GRCh38 \
  -m metadata.tsv \
  --sample-name genotype,stim,donor,replicate \
  --sample-type genotype,stim \
  --de-columns genotype,stim \
  --de-block donor

The default full workflow runs through step 12:

1. preprocessing and alignment (steps 1-4)
2. filtered BAMs and alignment QC (steps 5-7)
3. stranded BigWigs and hubs (steps 8-10)
4. featureCounts table (step 11)
5. DESeq2 analysis (step 12)

STAR and STAR-TE are available with -M. STAR-TE is a lab-specific STAR preset that permits up to 100 alignments per read (--outFilterMultimapNmax 100 and --winAnchorMultimapNmax 100) so repetitive-element reads are retained for compatible downstream methods. It does not quantify transposable elements by itself, and most conventional gene-level RNA-seq analyses should use HISAT2 or standard STAR instead.

Use --remove-duplicates only when the experimental design specifically requires duplicate removal; retaining RNA duplicates is the assay-aware default.

For storage-constrained projects, --retention-policy pruned preserves reusable filtered BAMs while removing bulky upstream intermediates after they are safe to delete.