Skip to content
PaperFren

Research method

RNA Sequencing (RNA-seq)

20 studies

Ask PaperFrenWatch this

RNA sequencing inventories RNA molecules in a sample by converting them to a cDNA library and counting short reads. Depending on the protocol it can quantify gene expression, assemble a transcriptome de novo, or detect non-canonical RNAs such as circular back-splice junctions. The output is a table of counts or assembled transcripts, not a proof that any one transcript causes the phenotype that motivated the experiment.

Biologists reach for RNA-seq when they need an unbiased picture of what a tissue, infection, or mixed culture is expressing rather than a handful of qPCR targets. It answers 'which transcripts differ, and by how much, under this contrast?' Its main limitation is that differential expression and homology-based pathway labels are not functional proof, and library chemistry (poly(A) versus total RNA, mixed-species RNA) changes what can be seen.

Studies

20

Findings

5

8 supporting · 0 challenging · 0 qualifying citations

Open tensions

2

Currently

What we know

  1. Protocol choice decides which molecules even appear. A computational pipeline on non-poly(A) RNA-seq was required to catalogue mammalian circular RNAs via back-splice junctions; most confident circRNAs used canonical GT-AG splice signals, and circular fractions varied by locus and cell type. A standard poly(A) mRNA-seq experiment would have missed them.
  2. De novo assemblies turn RNA-seq into a catalogue for species without a finished genome. Illumina sequencing of Aloe vera root and leaf produced on the order of 43,000 CDS per tissue with annotations tied to secondary-metabolite pathways; bacteria-challenged sea bass head kidney and spleen yielded a large immune transcriptome including hepcidin, lysozyme and RAG annotations.
  3. Time-of-day and mixed-culture design change the biological contrast RNA-seq is actually making. Tomato–DC3000 resistance was highest at 08:00 and susceptibility at 20:00, with red light the most disease-suppressive wavelength among those tested; in a chemostat co-culture, L. thermotolerans contributed about 24% of transcripts anaerobically versus about 8% aerobically.
  4. RNA-seq can be joined to genetics and to metabolites, still without proving causation. In 77 Yoruba LCLs, inter-individual methylation tracked genotypes and correlated with RNA-seq expression; in Panax ginseng, MEP-pathway expression patterns were taken to contribute to ginsenoside biosynthesis with IspD predicted as a key enzyme.
  5. Field disease RNA-seq separates states more cleanly than it names a pathogen. Hundreds of DEGs separated white-syndrome coral tissue from healthy tissue, while adjacent and healthy transcriptomes were more similar — expression differences, not a settled etiology.

Largest unresolved question

Reference-guided counting and de novo assembly are not interchangeable products. CircRNA detection needed non-poly(A) libraries and back-splice calling; Aloe and sea bass papers assemble transcripts because a complete reference is missing or incomplete; LCL and tomato papers quantify against known genes. Treating every RNA-seq paper as a gene-count table hides those design choices.

Procedure

How it works

The main stages of the method. Specific protocols and analysis choices vary between studies.

  1. Extract RNA from the sample

    Total RNA is purified from a tissue, culture, or mixed community. Integrity and which cells entered the tube bound what can later be counted; a degraded or biased extract cannot be repaired by deeper sequencing.

  2. Build a cDNA library

    RNA is reverse-transcribed and fragmented into a sequencing library. Poly(A) selection, rRNA depletion, or a circular-RNA pipeline each keep a different subset; the chemistry is a filter, not a complete inventory of cellular RNA.

  3. Sequence short reads

    The library is sequenced as millions of short reads. Depth decides which rare transcripts are seen; in mixed-species RNA the scarcer partner can drop below detection unless its share is high enough.

  4. Align or assemble, then count

    Reads are mapped to a genome or assembled de novo and summarised as gene counts or transcript models. Homology labels on those models are annotations, not enzyme assays.

  5. Call differential expression

    A statistical contrast reports which transcripts differ under the experimental design. Multiple-testing correction and batch structure matter here; an uncorrected list of 'up' genes is not a pathway proof.

  6. Treat the table as a map of transcripts, not a cause

    RNA-seq answers which transcripts differ, and by how much, under this contrast and this library chemistry. Differential expression in diseased tissue does not identify the cause of the disease, and a named pathway in the assembly has not been shown to run in the organism.

Common misconceptions

  • RNA-seq measures all RNA in a cell equally.

    Library chemistry filters the pool. Poly(A) selection misses many circular RNAs that a non-poly(A) pipeline was built to find, and mixed-species co-cultures only recover the scarcer partner when its transcript share is high enough (about 24% anaerobic versus 8% aerobic for L. thermotolerans).

  • If a pathway is annotated in the assembly, those enzymes have been shown to run that pathway in the organism.

    Aloe vera CDS counts are homology annotations, not enzyme assays; ginseng MEP conclusions still need functional tests; sea bass immune labels are homology-based, not a vaccine trial.

  • Differentially expressed genes in diseased tissue identify the cause of the disease.

    Coral white-syndrome RNA-seq shows hundreds of DEGs and that adjacent tissue looks more like healthy tissue; the authors do not treat that as proven causation, and field etiology remains debated.

Related