1. Spliced Read Alignment to Reference Plant Genomes

Eukaryotic RNA-Seq reads derived from mature mRNA span spliced exon-exon junctions. Standard genomic aligners (e.g. BWA-MEM, Bowtie2) cannot map junction reads without introducing massive mismatch penalties. We utilize HISAT2 (Hierarchical Indexing for Spliced Alignment of Transcripts) or STAR, incorporating known gene models from GFF3/GTF files.

2. Building the HISAT2 Plant Splice-Site Aware Index

#!/usr/bin/env bash
# Extract splice sites and exons from Solanum tuberosum GFF3/GTF
hisat2_extract_splice_sites.py Solanum_tuberosum.DM_v6.1.gtf > potato_splicesites.txt
hisat2_extract_exons.py Solanum_tuberosum.DM_v6.1.gtf > potato_exons.txt

# Build indexed reference (incorporating splice sites and exons into Graph FM-index)
hisat2-build -p 16 \
    --ss potato_splicesites.txt \
    --exon potato_exons.txt \
    Solanum_tuberosum.DM_v6.1.dna.toplevel.fa \
    hisat2_index/potato_v6_index

3. Executing Spliced Alignments with Downstream Transcript Assembly Flags

Crucially, when running HISAT2 for transcriptome quantification, the --dta (Downstream Transcript Assembly) flag must be specified to tailor alignment scoring for tools like featureCounts, HTSeq, and StringTie.

#!/usr/bin/env bash
mkdir -p bam_alignments

# Align paired-end reads and generate coordinate-sorted BAM files
for R1 in trimmed_fastq/*_clean_R1.fastq.gz; do
    R2="${R1/_clean_R1.fastq.gz/_clean_R2.fastq.gz}"
    SAMPLE=$(basename "$R1" _clean_R1.fastq.gz)
    
    echo "Aligning ${SAMPLE} to Solanum tuberosum DM1-3 v6.1..."
    
    hisat2 -p 12 \
        --dta \
        -x hisat2_index/potato_v6_index \
        -1 "$R1" \
        -2 "$R2" \
        --summary-file "bam_alignments/${SAMPLE}_summary.txt" | \
    samtools view -@ 4 -bS - | \
    samtools sort -@ 4 -o "bam_alignments/${SAMPLE}.sorted.bam" -
    
    # Generate index (.bai)
    samtools index "bam_alignments/${SAMPLE}.sorted.bam"
done