StudyRareStudyRare

RNA Sequencing (RNA-seq)

Log in to star

Last updated Nov 10, 2026

Log in to add personal notes on this page.

RNA sequencing measures the transcripts a gene actually makes. DNA sequencing shows what a variant is; RNA shows what it does to the message: whether a splice site is still used, whether an exon is skipped or an intron kept, and whether one allele's transcript is missing. In clinical genetics its main uses are settling variants of uncertain significance that may affect splicing and finding diagnoses that exome or genome sequencing left unsolved. In cancer, RNA is the preferred material for detecting gene fusions.

  • From RNA to reads: Reverse transcriptase copies RNA into complementary DNA (cDNA), which is then sequenced. A targeted test (RT-PCR of one gene, see PCR) answers a question about one variant; transcriptome-wide RNA-seq looks at every expressed gene at once.
  • Aberrant splicing: Reads that join exons in an unexpected way show exon skipping, intron retention, use of a cryptic splice site, or inclusion of a pseudoexon from deep in an intron.
  • Expression outliers: A gene expressed far below the level in control samples points to a variant that blocks transcription or makes an unstable transcript.
  • Allele-specific expression: When a heterozygous SNP in the transcript shows only one allele, the other allele's transcript is missing. That suggests a variant on the silent allele, even before the variant itself is found.
  • Nonsense-mediated decay (NMD): Cells destroy transcripts that carry a premature stop codon, so a variant that triggers NMD may leave little abnormal transcript to see. Labs can grow fibroblasts with an NMD inhibitor such as cycloheximide to capture it.
  • Tissue choice: A gene can be studied only in a tissue that expresses it. Blood is easy to collect but expresses fewer disease genes; cultured skin fibroblasts express more; a muscle biopsy is used for muscle disease. In published cohorts of patients still undiagnosed after genetic testing, RNA-seq found a diagnosis in about 8% using blood and about 35% using muscle in muscle disease.
  • Sample handling: RNA degrades within hours, so blood is drawn into a tube that stabilizes it (such as PAXgene).
  • Splice-site variants of uncertain significance: A variant near a splice site, or a deep intronic variant flagged by a prediction tool such as SpliceAI, can be tested directly. Under the ClinGen splicing recommendations (2023), RNA showing that a variant makes loss-of-function transcripts is scored as PVS1 (RNA), at the strength the PVS1 decision tree assigns. An in-frame change may be downgraded depending on what the lost region does, and a variant that still leaves some normal transcript is judged case by case and may not count toward pathogenicity. PS3 is not added on top for the same RNA result. RNA showing normal splicing for an intronic or synonymous variant supports BP7 at strong level (BP7_Strong (RNA)). See ACMG variant classification.
  • Unsolved exome or genome: RNA-seq can show what a variant already found in DNA does, such as a deep intronic variant that adds a pseudoexon, and its expression outliers point to genes the DNA analysis passed over.
  • A missing second hit: A proband with one pathogenic variant in a recessive gene that fits the phenotype may carry a second, noncoding variant. Loss of expression from the other allele is evidence for it.
  • Cancer gene fusions: A fusion joins two genes, often with breakpoints inside large introns that DNA panels cover poorly. RNA reads the joined transcript directly, so it catches fusions DNA testing misses. RNA-based NGS is the preferred test for NTRK1, NTRK2, and NTRK3 fusions, which predict response to TRK inhibitors. ALK fusions in lung cancer are usually found by immunohistochemistry, FISH, or a DNA panel, with RNA testing adding cases those miss. See Somatic tumor variants.
  • The gene must be expressed in the sample: Many neurologic and cardiac genes are barely expressed in blood or fibroblasts, so a normal result in the wrong tissue does not exclude a splicing effect.
  • NMD hides abnormal transcripts: Without an NMD inhibitor, a truncated transcript may be invisible.
  • Partial effects are hard to judge: A variant that leaves some normal transcript may or may not cause disease; the answer depends on how much of the normal product the gene needs.
  • Tissue-specific splicing: Splicing in blood or fibroblasts may differ from splicing in the affected organ.
  • Availability: Fewer clinical labs offer transcriptome-wide RNA-seq than exome or genome sequencing; it needs its own sample and a large set of normal controls for comparison.
  • DNA shows that a variant sits near a splice site; RNA shows whether splicing actually changed.
  • Send RNA from a tissue that expresses the gene, because a gene that is silent in the sample gives no answer.
  • When DNA testing finds no fusion in a tumor where one is suspected, RNA is the next sample, because the fused transcript is read directly.