bulk-rnaseq

To build a bulk RNA-seq pipeline, you validate the samplesheet, run QC and trimming, perform alignment or pseudoalignment, assemble a gene-level counts matrix, and run differential expression using PyDESeq2, followed by GSEA enrichment.

Yes, Salmon pseudoalignment output is supported. The pipeline converts Salmon quantification results into a gene-by-sample integer counts matrix using pytximport, which then feeds directly into PyDESeq2 for differential expression analysis.

Yes, you need a metadata template aligned with your counts matrix. The pipeline checks replication, batch structure, and strandedness to ensure your metadata is correctly formatted for PyDESeq2 differential expression and subsequent GSEA enrichment handoff.

The best way to convert STAR or featureCounts output is using the pipeline's counts matrix bridge, which transforms raw alignment counts into a structured gene-by-sample matrix with paired metadata, automatically checking for common RNA-seq pitfalls.

You handle strandedness and batch structure by using the pipeline's design and QC guardrails, which validate replication, detect batch effects, and verify strandedness parameters before passing data to PyDESeq2 to ensure trustworthy differential expression results.

Integer gene-level counts are required because PyDESeq2 expects raw count matrices for accurate statistical modeling of read counts. The pipeline ensures counts are properly assembled and strandedness is handled before differential expression analysis.