Jaber H. Jaradat, MD
Bioinformatics RNA-Seq Translational Oncology R

From FastQC to Pathway Enrichment: A Primer on RNA-Seq for Clinicians

A clinical and computational walkthrough of high-throughput bulk and single-cell RNA-sequencing pipelines for translational oncology.

Dr. Jaber Husam Jaradat Jaber Husam Jaradat, M.D.
• • 6 min read

From FastQC to Pathway Enrichment: A Primer on RNA-Seq for Clinicians

As modern oncology moves rapidly toward precision biomarkers and targeted molecular therapies, understanding transcriptomic architectures is no longer reserved solely for dry-lab bioinformaticians. As clinicians, having hands-on fluency in how raw sequencing reads translate into clinically actionable gene signatures empowers us to bridge patient care with discovery.

In this guide, we review the fundamental stages of an RNA-seq workflow, highlighting practical considerations when working with solid tumor specimens.


1. Quality Assessment & Trimming

High-throughput Illumina sequencers generate raw reads in FASTQ format. Before aligning reads to the reference genome (e.g., GRCh38/hg38), thorough quality control is mandatory:

  • FastQC & MultiQC: Evaluate per-base sequence quality scores (Phred score $\ge 30$), GC content distributions, and adapter contamination.
  • Trimmomatic / Cutadapt: Remove low-quality base tails and residual adapter sequences.
# Example FastQC quality check
fastqc raw_sample_R1.fastq.gz raw_sample_R2.fastq.gz -o qc_reports/
multiqc qc_reports/ -o aggregated_qc/

2. Alignment & Quantification

Once cleaned, reads must be accurately mapped to the reference transcriptome:

  1. Splice-Aware Aligners (e.g., STAR, HISAT2): Essential for eukaryotic RNA due to splicing across exon-exon junctions.
  2. Pseudoaligners (e.g., Salmon, Kallisto): Exceptionally fast, alignment-free quantification of transcript abundances directly from transcriptomes.
# Alignment with STAR
STAR --runThreadN 8 \
     --genomeDir /ref/STAR_GRCh38 \
     --readFilesIn sample_trimmed_R1.fq.gz sample_trimmed_R2.fq.gz \
     --readFilesCommand zcat \
     --outSAMtype BAM SortedByCoordinate \
     --outFileNamePrefix ./aligned/sample_

3. Differential Expression Analysis in R

With count matrices in hand, statistical modeling accounts for library size differences, sequencing depth, and composition biases.

  • DESeq2 and edgeR utilize negative binomial generalized linear models (GLMs) to model count variability across biological replicates.
  • Variance Stabilizing Transformation (VST) prepares high-dimensional counts for PCA and hierarchical clustering.
library(DESeq2)

# Construct DESeqDataSet
dds <- DESeqDataSetFromMatrix(countData = counts,
                              colData = colData,
                              design = ~ condition)

# Run differential testing
dds <- DESeq(dds)
res <- results(dds, contrast=c("condition", "Tumor", "Normal"))
head(res[order(res$padj), ])

4. Functional Interpretation & Pathway Enrichment

Identifying a list of 500 differentially expressed genes is just the starting point. Translating genes into biological mechanisms requires enrichment against curated ontologies:

  • Over-Representation Analysis (ORA): Tests whether a specific pathway (e.g., KEGG, Reactome, GO) has more DEGs than expected by chance.
  • Gene Set Enrichment Analysis (GSEA): Evaluates ranked gene lists across hallmark cancer signatures without arbitrary p-value cutoffs.

Key Takeaway for the Clinic

Bioinformatics is not a disconnected dry-lab exercise. When paired with tumor biology and clinical outcomes, transcriptomic pipelines unlock actionable insights into drug resistance, immune evasion, and patient prognosis.

Dr. Jaber Husam Jaradat

About Dr. Jaber Husam Jaradat

Physician-scientist and medical intern pursuing U.S. Internal Medicine training with a focus in cancer bioinformatics, single-cell/spatial omics, and translational immuno-oncology. Recipient of the 2026 AACR Global Scholar-in-Training Award (GSITA) and founder of BrighterUs.