Master the complete RNA-Seq transcriptomics workflow from raw reads to biological interpretation, with two execution pathways for model organisms (reference alignment and alignment-free quantification) and a dedicated pipeline for non-model organisms using de novo transcriptome assembly.

1) Workflow Foundations & Project Setup — convert biological question to an analysis design you can defend
• Study design: condition labels, biological replicates, contrasts, pairing, covariates
• Metadata: sample sheet schema (sample_id, condition, batch, lane, library_type, strandedness)
• Reference selection: genome build + GTF/GFF, transcriptome FASTA (consistency matters)
• Folder structure: raw/trimmed/aligned/counts/qc/results/reports + naming conventions
• Reproducibility: tool versions, parameters, run logs, checksums (MD5/SHA)
Design Matrix Metadata Sheet Runbook Template
2) Raw Read Quality Control (FASTQ) — identify problems before you “waste compute” downstream
• Phred scores, per-base quality, per-sequence quality, N content
• Adapter contamination, overrepresented sequences, k-mer bias
• GC distribution shifts; duplication levels; read length distribution
• MultiQC aggregation: multi-sample dashboards for fast comparisons
FastQC MultiQC seqtk
QC Gate: document issues + decide trimming/filters before moving forward.
3) Trimming & Read Cleanup — remove technical artifacts while preserving biological signal
• Adapter trimming strategies; quality trimming (sliding window vs fixed cutoffs)
• Minimum length thresholds; handling orphan reads in paired-end
• Post-trim QC to verify improvement (before vs after)
• Common mistakes: over-trimming, inconsistent paired-end outputs, wrong adapter set
Trimmomatic Cutadapt fastp
QC Gate: adapter content reduced + acceptable read retention.
4A) Model Organism Lane — Reference-Based Alignment
• Genome indexing and splice-aware alignment concepts
• HISAT2 alignment: parameters that matter (paired-end, sensitivity, splicing)
• BAM processing: sorting/indexing + alignment summary interpretation
• Mapping QC: alignment rate, multi-mapping, insert size trends, strandness checks
HISAT2 samtools RSeQC (QC)
QC Gate: mapping rate + strandness consistent with library prep.
4B) Model Organism Lane — Alignment-Free Quantification (Alternative Route)
• When to choose pseudoalignment (speed, transcript-level quantification)
• Transcriptome indexing; abundance estimation; bootstraps (Kallisto)
• Outputs: TPM, estimated counts; merging across samples
• Common pitfalls: wrong transcriptome version, annotation mismatch, inflated isoforms
Salmon Kallisto tximport
QC Gate: assignment rate acceptable + no obvious reference mismatch.
5) Quantification for DE — Gene Counts (featureCounts)
• Gene model concepts: exons, genes, transcripts; how GTF drives counting
• Strandedness and feature type selection; paired-end counting choices
• Count matrix construction and hygiene (sample naming, gene ID consistency)
• Exploratory analysis: sample clustering, PCA/MDS, outlier detection
featureCounts GTF/GFF Count Matrix
QC Gate: replicates cluster together; obvious outliers explained.
6) Differential Expression (DE) — statistically correct results you can interpret
• Normalization concepts; dispersion; model design; contrasts
• Multiple testing correction (FDR) and effect size interpretation (log2FC)
• Plots: MA, volcano, heatmaps, PCA/MDS; result filtering best practices
• Result sanity checks: known markers, directionality, replicate consistency
DESeq2 edgeR limma-voom
QC Gate: DE results stable under reasonable thresholds; plots consistent.
7) Functional Interpretation & Enrichment (clusterProfiler)
• Gene ID mapping (ENSEMBL/ENTREZ) and gene universe selection
• ORA vs GSEA — when to use each; common biases and controls
• GO/KEGG/Reactome enrichment; dotplots and reporting structure
• How to write an interpretation narrative for results + limitations
clusterProfiler GO / KEGG / Reactome MSigDB
QC Gate: correct gene universe + correct IDs + biologically plausible themes.
8) Non-Model Organism Lane — De novo Transcriptome Pipeline
• Trinity assembly: compute planning, memory/CPU considerations, output structure
• De-redundancy: CD-HIT clustering to reduce transcript inflation
• Assembly assessment: Trinity-stats (N50, contig distribution) , read representation
• BUSCO: completeness and duplication interpretation; what “good” means
• Quantification: Kallisto quant on assembled transcriptome → DE workflow
Trinity CD-HIT Trinity-stats BUSCO Kallisto
QC Gate: BUSCO completeness acceptable; duplication explained; stable quant.
Deliverables (What You Take Home)
• QC pack: FastQC + MultiQC reports, trimming decisions, acceptance notes
• Processed data: trimmed FASTQs, aligned BAMs (model lane) , count/TPM matrices
• DE pack: DE tables, MA/volcano/heatmaps, sample clustering (PCA/MDS)
• Enrichment pack: GO/KEGG/Reactome tables + dotplots + interpretation template
• Non-model pack (Tier 3 emphasis) : assembled transcriptome + CD-HIT + Trinity-stats + BUSCO report
• Reproducibility pack: runbook (commands + parameters + versions) + folder template
What’s Included in Each Tier
Coverage	Tier 1 (Basic)	Tier 2 (Pro)	Tier 3 (Enterprise)
Model Lane (Alignment)	Full practical	Full practical + QC gates + troubleshooting	Full practical + team SOP + audit checklist
Alignment-Free Lane	Demo + interpretation	Full practical + tximport + comparisons	Full practical + standardization for teams
Non-Model (De novo)	Concept overview	Guided demo + QC interpretation	Full practical: Trinity → CD-HIT → BUSCO → Kallisto → DE
Capstone / Report	Templates + guided structure	Publication-ready plots + narrative framework	Team delivery pack + one dataset review
Apply Tier 1 Apply Tier 2 Apply Tier 3

FASTQ structure, quality scores & initial QC (MultiQC)
FastQC MultiQC seqtk
Adapter/quality trimming & contamination screening
Cutadapt fastp FastQ Screen
Design matrices, metadata & sample randomization
Theory

Splice-aware alignment & count generation
STAR HISAT2 featureCounts
Alignment-free quantification & transcript-level summaries
Salmon Kallisto tximport
Differential expression pipelines & visualization
DESeq2 edgeR limma-voom

Batch correction, QC metrics & exploratory analysis
sva/ComBat PCA/UMAP RSeQC
Differential splicing & isoform usage
DEXSeq rMATS SUPPA2
Microarray preprocessing & DE analysis (RMA/limma)
oligo/affy RMA limma

GSEA/ORA & pathway interpretation (KEGG/Reactome/GO)
clusterProfiler MSigDB Enrichr
End-to-end automation with provenance & dashboards
Snakemake R Markdown/Quarto pandas
Publication-ready figures: MA, volcano, heatmaps, PCA
ggplot2 ComplexHeatmap EnhancedVolcano