Module 2: Experimental Design, Library Preparation Strategies & Sequencing Depth
Module 2: Experimental Design, Library Preparation Strategies & Sequencing Depth
Designing rigorous RNA-seq experiments: biological vs. technical replicates, power analysis, Poly(A) mRNA selection vs. Ribo-Zero ribosomal RNA depletion, strandedness (dUTP second-strand), and target coverage.
2. RNA-Seq Experimental Design, Library Prep & Sequencing Depth
A successful transcriptomics study begins long before sequencing data is generated. Flawed experimental designs—such as confounding batch effects or lack of biological replicates—cannot be rescued by downstream bioinformatics.
Biological Replicates vs. Technical Replicates
In high-throughput sequencing, technical variance (variation between different lanes or flow cells for the exact same cDNA library) is exceptionally low. In contrast, biological variance (variation between independent cell cultures, animal specimens, or patient biopsies) is substantial.
- Minimum Gold Standard: $\ge 3$ biological replicates per condition are required for cell lines; $\ge 5\text{–}10$ replicates for inbred animal models; and $\ge 20\text{–}100$ for human clinical cohorts.
- Statistical Consequence: Replicates allow DESeq2/edgeR to accurately estimate gene-specific dispersion parameters. With $N=1$ or $N=2$, reliable statistical hypothesis testing is impossible.
Library Preparation Strategies: Poly(A) vs. Ribo-Zero
Poly(A) mRNA Enrichment (Oligo-dT)
Magnetic beads coated with oligo-dT capture mature eukaryotic mRNAs containing polyadenylated 3' tails.
Pros: Highly cost-effective; maximizes sequencing depth on protein-coding genes.
Cons: Ineffective on degraded RNA (FFPE samples with RIN < 6); excludes non-polyadenylated long non-coding RNAs (lncRNAs) and bacterial transcripts.
Ribosomal RNA Depletion (Ribo-Zero / Total RNA)
Biotinylated antisense probes hybridize to cytoplasmic (18S, 28S) and mitochondrial (12S, 16S) rRNAs (>85% of total RNA), which are removed via streptavidin beads.
Pros: Ideal for degraded clinical FFPE samples; captures non-coding RNAs, nascent pre-mRNAs, and bacterial transcripts.
Cons: Requires 1.5× to 2× higher sequencing depth to achieve equivalent coding gene coverage.
Strand-Specific (Stranded) Sequencing
Modern protocols preserve the original orientation of RNA transcripts (dUTP second-strand marking with actinomycin D). Stranded RNA-seq distinguishes overlapping sense and antisense transcripts, prevents ambiguous read assignment in dense eukaryotic genomes, and improves quantification accuracy.
Common Pitfalls & Troubleshooting Checklist
- Java Heap Exhaustion (GATK / Picard): If processes terminate with
java.lang.OutOfMemoryError, set your environment variable explicitly:export _JAVA_OPTIONS="-Xmx16g -Djava.io.tmpdir=/tmp". - Index Timestamp Mismatch: If tools complain that a BAM index is missing or out of date, regenerate with
samtools index -@ 4 sample.bam. - Conda Channel Conflicts: Avoid ABI incompatibilities between bioconda and conda-forge by setting strict priority:
conda config --set channel_priority strict. - Chromosome Naming Contamination: Always verify whether reference FASTA headers use
chr1or1; divergent naming causes silent alignment drops.