Chapter 1: Next-Generation Sequencing Technology: Types, Principles & RNA-Seq File Formats

1. Foundations of Next-Generation Sequencing in Transcriptomics

Transcriptomics quantifies the complete set of RNA transcripts produced by the genome under specific physiological conditions, developmental stages, or therapeutic perturbations. Understanding gene expression begins at the physical sequencing flowcell. Modern high-throughput sequencing converts microscopic cDNA or RNA molecules into digital nucleotide reads through four dominant technological paradigms:

2. RNA Library Preparation: Poly(A)+ vs. Ribo-Zero vs. Small RNA

Total human RNA consists of roughly 80–90% ribosomal RNA (28S, 18S, 5.8S, and 5S rRNA), 10–15% transfer RNA (tRNA), and only 1–4% protein-coding messenger RNA (mRNA). Sequencing total RNA without depletion or enrichment would result in over 90% of sequencing reads aligning uselessly to rRNA operons. Two primary library preparation strategies dominate biomedical research:

3. RNA-Seq File Architecture: From FASTQ to Digital Count Matrices

Processing transcriptomic data requires transitioning through four standardized file formats. Each format represents a distinct milestone in the computational pipeline:

# The 4 Core Stages and File Formats of Biomedical RNA-Seq
[ 1. Raw Reads ]      FASTQ.GZ    (Base calls + Phred quality scores Q = -10 * log10(P))
       │
       ▼ (Alignment via STAR / HISAT2 or Quasi-mapping via Salmon / Kallisto)
[ 2. Alignments ]     BAM / SAM   (Genomic coordinates, CIGAR strings, spliced junction tags)
       │
       ▼ (Quantification against GTF/GFF3 gene models via featureCounts / HTSeq)
[ 3. Annotations ]    GTF / GFF3  (Exon boundaries, transcript IDs, gene biotypes)
       │
       ▼ (Aggregation into statistical modeling matrix)
[ 4. Count Matrix ]   TSV / RDS   (Rows = Gene Symbols / Ensembl IDs; Columns = Patient Samples)

The Four-Line Anatomy of a FASTQ Entry

Each sequenced read in a FASTQ file is represented by exactly four lines:

@A00123:456:H5WNYDSX2:1:1101:1020:1000 1:N:0:ATCACG
NATCGATCGATCGATCGATCGTAGCTAGCTAGCTAGCTAGCTAGCTAGCTAAGCTAGCTA
+
#FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF

• Line 1 (Header): Instrument ID (@A00123), Run number (456), Flowcell ID (H5WNYDSX2), Lane (1), Tile (1101), X/Y coordinates (1020:1000), Read pair (1), Filter flag (N = passed), and Multiplex Index barcode (ATCACG).
• Line 2 (Sequence): Raw base calls (A, C, T, G, and N for ambiguous bases).
• Line 3 (Separator): Plus sign (+), optionally repeating the header.
• Line 4 (Quality Scores): ASCII characters representing Phred quality scores where character I represents Q40 (1 error in 10,000 bases) and # represents Q2 (high probability of error).

Practical Command-Line: Verifying Strandedness and Read Metrics

# 1. Inspect first 2 reads from paired-end FASTQ files
zcat sample_tumor_R1.fastq.gz | head -n 8

# 2. Check total read count in gzipped FASTQ file
echo $(zcat sample_tumor_R1.fastq.gz | wc -l) / 4 | bc

# 3. Infer strandedness (Forward, Reverse, or Unstranded) using RSeQC or check_strandedness
# In modern dUTP / TruSeq Stranded protocols, reads are almost universally Reverse (RF / fr-firststrand)
check_strandedness -g Homo_sapiens.GRCh38.112.gtf \
                   -r1 sample_tumor_R1.fastq.gz \
                   -r2 sample_tumor_R2.fastq.gz