10: Bulk RNA-Seq Alignment & Quantification

Chapter 10: Bulk RNA-Seq Alignment & Quantification

Transcriptome profiling via RNA Sequencing (RNA-seq) quantifies the dynamic expression of tens of thousands of genes across conditions, tissues, and treatments. Unlike DNA sequencing, mature eukaryotic mRNA molecules are spliced: non-contiguous exons are joined together after intron excision. Consequently, RNA-seq reads cannot be mapped with standard DNA aligners like BWA-MEM; they require specialized splice-aware aligners or alignment-free pseudoaligners.

1. Spliced Alignment with STAR (2-Pass Mode)

# Run STAR in 2-Pass mode for superior novel splice junction discovery
STAR \
    --runThreadN 16 \
    --genomeDir /databases/star_grch38_index \
    --readFilesIn rna_sample_R1.fastq.gz rna_sample_R2.fastq.gz \
    --readFilesCommand zcat \
    --outFileNamePrefix rna_sample_ \
    --outSAMtype BAM SortedByCoordinate \
    --twopassMode Basic \
    --quantMode GeneCounts

2. Ultra-Fast Pseudoalignment with Salmon

Salmon achieves blistering speed (<5 minutes per sample) by bypassing base-to-base alignment. It streams reads against a transcriptome index, corrects for sequence-specific GC biases, and performs expectation-maximization (EM) optimization:

# Quantify transcripts directly using Salmon with GC and sequence bias correction
salmon quant \
    -i /databases/salmon_grch38_index \
    -l A \
    -1 rna_sample_R1.fastq.gz \
    -2 rna_sample_R2.fastq.gz \
    --validateMappings \
    --gcBias \
    --seqBias \
    -p 16 \
    -o salmon_quant_output/