The transcriptome is the complete set of RNA transcripts present in a cell, tissue, or organism at a particular time and under particular conditions. It includes protein-coding and non-protein-coding RNAs and can be described in terms of their sequences, molecular structures, and abundances. Unlike the genome, which is relatively stable within a cell lineage, the transcriptome changes with development, environmental conditions, physiological activity, and disease. Its systematic study is called transcriptomics.
Scope and composition
A transcript is an RNA molecule produced by transcription from a DNA template, although the RNA molecules measured in a sample may have undergone extensive processing after transcription. “Transcriptome” can refer either to all RNA molecules in a biological system or to the collection of distinct transcript species detected there. Consequently, transcriptome studies must specify both their biological scope and the RNA classes captured by their methods.
Major components include:
- Messenger RNA (mRNA): transcripts that carry information used to synthesize proteins.
- Ribosomal RNA (rRNA): structural and catalytic components of ribosomes, accounting for most cellular RNA by mass in many actively growing cells.
- Transfer RNA (tRNA): adaptor molecules that connect amino acids with codons during protein synthesis.
- Non-coding RNA: a broad category that includes rRNA and tRNA, as well as long non-coding RNAs, microRNAs, small nuclear RNAs, and other regulatory or processing-associated molecules.
- Precursor and incompletely processed RNAs: molecules that retain sequences subsequently removed during RNA maturation.
In eukaryotes, a single gene can give rise to multiple transcript isoforms through alternative transcription initiation, alternative splicing, and alternative sites of RNA cleavage and polyadenylation. Transcriptomes can also contain RNA from organellar genomes, including those of mitochondria and chloroplasts.
Many experiments measure only a subset of this diversity. For example, enrichment for polyadenylated RNA captures many eukaryotic mRNAs but excludes most rRNA and tRNA and misses numerous non-polyadenylated transcripts. Such an experiment is often described as transcriptome profiling, even though it does not survey all RNA classes.
Biological meaning and dynamics
The transcriptome provides a molecular view of gene expression. Cells with broadly similar DNA sequences can have very different transcriptomes because they express different genes and regulate RNA processing and degradation differently. These differences contribute to cell identity and changing functional states.
Transcript abundance is not determined by transcription alone. It reflects the balance between RNA production and removal, together with processing, transport, and dilution during cell growth. A simplified model for the abundance of an RNA species is
where is its synthesis rate and is a first-order removal constant. Under steady-state assumptions,
Thus, an increase in RNA abundance can result from increased synthesis, increased stability, or both. Conventional abundance measurements do not, by themselves, distinguish these possibilities.
Transcriptomes vary over developmental time, across the cell cycle, and in response to nutrients, temperature, signaling molecules, and other conditions. RNA abundance can also fluctuate among otherwise similar cells because of intermittent transcription and differences in regulatory state.
A transcriptome is not equivalent to a proteome. Translation, protein degradation, and post-translational regulation affect the relationship between RNA abundance and protein abundance. Transcriptomics and proteomics therefore provide complementary rather than interchangeable information.
Development of transcriptome research
Early studies of gene expression generally examined individual transcripts or relatively small sets of genes. Large-scale complementary DNA sequencing, expressed sequence tags, and serial analysis of gene expression expanded the number of transcripts that could be surveyed.
DNA microarrays subsequently enabled parallel measurements of thousands of transcripts by hybridizing labeled sample material to predefined probes. They made genome-scale expression comparisons widely accessible, but their coverage depended on probe design and prior knowledge of transcript sequences.
From the late 2000s, RNA sequencing (RNA-seq) increasingly became the principal approach for broad transcriptome analysis. Sequencing allowed researchers to measure expression while also identifying previously unannotated transcripts, splice junctions, and sequence variation. Later developments expanded transcriptomics to individual cells, isolated nuclei, spatially organized tissue samples, and long RNA molecules.
These technologies differ in what they capture. Greater throughput does not necessarily imply more complete coverage of every RNA class or more accurate reconstruction of every transcript.
Methods of measurement
Bulk RNA sequencing
Bulk RNA-seq profiles RNA extracted from a population of cells. A typical workflow includes RNA extraction, quality assessment, selection or depletion of particular RNA classes, library preparation, sequencing, and computational analysis.
Most workflows convert RNA into complementary DNA before sequencing, although direct RNA sequencing is also available. Reads may be aligned to a reference genome or transcriptome, or assembled without a complete reference.
Bulk measurements are useful for comparing tissues or experimental conditions. However, they average over the cells in a sample. An apparent expression difference can arise because individual cells change their expression, because the proportions of cell types change, or because both occur.
Single-cell and single-nucleus profiling
Single-cell RNA sequencing assigns RNA measurements to individual cells, commonly using molecular barcodes. It can reveal cell populations and states that bulk measurements obscure.
Single-nucleus RNA sequencing instead profiles RNA from isolated nuclei. It is useful for some frozen or difficult-to-dissociate tissues, but nuclear RNA composition differs from that of whole cells. Nuclear measurements often contain a higher proportion of incompletely spliced transcripts.
Many high-throughput methods sample only transcript ends and capture a limited fraction of the molecules present. A zero count therefore does not necessarily mean that a gene was biologically inactive. Computational groupings of cells also require interpretation: they may represent stable cell types, transient states, technical effects, or mixtures of these.
Spatial transcriptomics
Spatial transcriptomics measures RNA while retaining information about its location in a tissue. Methods include sequencing-based capture on spatially indexed surfaces and imaging-based detection of RNA molecules.
Spatial resolution and transcript coverage vary across platforms. Some measurements combine signals from several cells; others localize individual RNA molecules. Spatial information helps connect gene expression with tissue organization, but molecular proximity alone does not demonstrate communication or a causal interaction between cells.
Complementary approaches
Long-read sequencing can help identify complete transcript structures and distinguish isoforms that are difficult to resolve with short reads. Small-RNA sequencing uses preparation methods suited to short regulatory RNAs. Metabolic RNA labeling and nascent-transcription assays investigate newly synthesized RNA rather than only accumulated abundance.
Targeted methods, including reverse-transcription quantitative polymerase chain reaction and in situ hybridization, are often used to examine selected findings with different measurement principles.
Analysis and interpretation
Transcriptome analysis commonly addresses several distinct questions:
- Expression quantification: estimating RNA abundance at the gene or transcript level.
- Differential expression: testing whether abundance differs between defined groups.
- Transcript reconstruction: identifying transcript structures, splice junctions, or previously unannotated RNAs.
- Isoform usage: determining whether the relative contributions of transcripts from the same gene change.
- Functional interpretation: examining whether groups of affected genes share biological functions.
- Cell-state analysis: identifying patterns of cellular similarity, variation, and response.
Sequencing counts depend on library size, capture efficiency, transcript properties, and the composition of the sample. Normalization is therefore necessary for meaningful comparisons. Measures such as transcripts per million account for transcript length and sequencing depth, but are not universal substitutes for the statistical models used in differential-expression testing.
Most sequencing measurements are relative rather than absolute. If one group of transcripts expands substantially, the measured proportions of other transcripts can fall even when their absolute molecule numbers remain unchanged. External controls or independent measurements can help investigate such effects, although they introduce their own assumptions.
Biological replication is essential for estimating variability. Thousands of measured cells from one organism do not automatically provide thousands of independent replicates for a comparison between organisms. Analyses must also address multiple testing, confounding variables, and batch effects.
Applications and limitations
Transcriptomics is used to investigate development, cell differentiation, environmental responses, host–microbe interactions, and the effects of genetic or experimental perturbations. In organisms with incomplete genome annotations, transcript data help identify expressed genes and improve gene models. Community-level RNA profiling, often called metatranscriptomics, examines expression across mixtures of microorganisms.
In biomedical research, transcriptomes are used to characterize disease-associated cellular states and identify candidate biomarkers. A reproducible association between an RNA pattern and a condition does not by itself establish a mechanism or demonstrate clinical usefulness.
Important limitations include RNA degradation during collection, changes induced by tissue dissociation, incomplete capture of low-abundance transcripts, and difficulty assigning reads to similar sequences. Transcript structure, chemical modifications, and library preparation can introduce additional biases. Reference annotations may omit relevant isoforms or represent some organisms and populations better than others.
Detection also does not establish biological function: a transcript may be short-lived, a by-product of transcription, or functional only under particular conditions. Conversely, failure to detect a transcript is not proof of its absence.
Finally, an ordinary transcriptome profile is a snapshot of RNA abundance, not a direct measurement of regulatory causation, protein activity, or metabolic flux. Establishing mechanisms usually requires additional evidence, such as controlled perturbations, time-resolved measurements, or independent molecular assays.