Transcription, Splicing, and RNA Degradation

Three Processes, Three Different Roles

Spliced and unspliced RNA counts describe RNA molecules at different processing stages, but a count is not itself a rate. To interpret RNA dynamics, distinguish the processes that produce, convert, and remove RNA. Transcription produces a precursor transcript; splicing removes intronic sequence and converts eligible unspliced RNA into spliced RNA; and degradation removes RNA molecules. These processes can affect different RNA pools and need not proceed at the same pace.

The introductory material on spliced and unspliced RNA counts establishes that sequencing-based measurements can be separated into these two categories. Here, the focus is the underlying biological interpretation of each process. For clarity, let \(U\) denote the number of unspliced molecules of a gene in a cell at a particular time, and let \(S\) denote its number of spliced molecules. These are molecular abundances in the biological description; observed sequencing counts are measurements related to those abundances, not guaranteed exact molecule counts.

A useful distinction is therefore between a rate constant and a flux. A transcription rate such as 12 molecules per minute already describes a production flux. By contrast, a splicing rate constant such as \(0.08\) per minute must be combined with the amount of eligible unspliced RNA to obtain a splicing flux. If there are \(U\) eligible molecules and each has the same first-order splicing rate constant \(\beta\), the expected instantaneous splicing flux is \(\beta U\) molecules per unit time.

Transcription Produces Precursor RNA

Transcription is the process by which RNA is synthesized from a DNA template. In the simplified picture used to interpret spliced and unspliced RNA, the immediate product is a precursor that contains intronic sequence and can contribute to the unspliced pool. The effective transcription rate, often denoted \(\alpha\), is expressed in molecules per unit time. It is a production rate, not a per-molecule rate constant: it does not need to be multiplied by the existing RNA abundance to determine the production flux.

This effective rate should not automatically be equated with promoter initiation frequency. A transcript must be initiated and synthesized, and experimental observations may reflect delays or incomplete transcripts. In a simplified kinetic description, \(\alpha\) summarizes the rate at which completed precursor molecules enter the RNA pool being considered. The interpretation depends on what counts as entry into that pool and on whether transcription is approximately stable over the time interval of interest.

Transcription can vary across genes, cells, and time. A larger unspliced count does not, by itself, prove that transcription is faster: the count could also be larger because splicing or degradation is slower, or because the cell was sampled during a transient change. Counts are snapshots of abundance, while transcription is a process rate.

Splicing Converts RNA Between Pools

Splicing removes intronic regions from a precursor transcript. In the two-pool description, it is a conversion: an eligible molecule leaves the unspliced pool and contributes to the spliced pool. It does not create a new RNA molecule from nothing, nor does it itself mean that the RNA has been degraded. Real transcripts may undergo multiple intron-removal events, so a single effective splicing rate can summarize a more complicated sequence of processing steps.

Under a first-order approximation, each molecule in the eligible unspliced pool has the same instantaneous chance of being spliced per unit time. If the rate constant is \(\beta\), with units of inverse time, and the pool contains \(U\) eligible molecules, then the expected instantaneous conversion flux is

The units confirm the interpretation: \((1/\text{time})\times(\text{molecules})=\text{molecules}/\text{time}\). This calculation depends on the first-order approximation and on \(U\) referring to molecules that are eligible for the modeled conversion. If processing is delayed, depends on transcript features, or occurs in multiple stages, one effective \(\beta\) may not capture all relevant biology.

Degradation Removes RNA

Degradation is the removal of RNA molecules. It can act on unspliced RNA, spliced RNA, or both, and the rates need not be equal. For example, a precursor that is incorrectly processed may be removed, while a mature transcript may have a different lifetime determined by its sequence, cellular context, and interactions with RNA-binding molecules. It is therefore often more biologically meaningful to distinguish degradation rates for different pools than to speak of one universal RNA degradation rate.

Under a first-order degradation assumption, each molecule in a specified pool has a constant instantaneous removal hazard. If \(\gamma\) is the degradation rate constant, its units are inverse time, and the expected instantaneous degradation flux from a pool of \(N\) molecules is \(\gamma N\) molecules per unit time. This assumption means the per-molecule removal rate is constant over the interval being modeled; it does not mean the pool size itself is constant.

A consequence of a constant first-order hazard is an exponential survival curve. If \(P(T>t)\) is the probability that an individual RNA molecule remains undegraded beyond time \(t\), then over a short interval \(\Delta t\), the probability of surviving is approximately multiplied by \(1-\gamma\Delta t\). Taking the continuous-time limit gives the survival function below. The half-life is the time at which this survival probability equals one half.

This derivation relies on a constant per-molecule hazard and a homogeneous interpretation of the pool. If molecules have different hazards, or the cellular environment changes during the measurement, a single exponential survival curve and a single half-life may be inadequate summaries.

Rate Assumptions, Measurements, and Failure Modes

These rate interpretations rely on specific assumptions. A first-order splicing or degradation constant assumes a proportional relationship between pool size and instantaneous flux, with the same effective per-molecule rate for molecules in the pool. A transcription rate expressed in molecules per time assumes a defined interval and a consistent definition of the precursor product. Each rate is local to its biological context: it may change with cell state, gene, time, or experimental condition.

Sequencing counts do not directly reveal these rate constants. Counts are affected by capture efficiency, sequencing depth, transcript features, and the probability that a molecule is assigned to the unspliced or spliced category. A count of zero can reflect low abundance, incomplete capture, or sampling variability; it is not proof that the corresponding biological process has stopped. Likewise, a large unspliced count is compatible with high transcription, slow splicing, slow degradation, or combinations of these mechanisms.

A common mistake is to interpret a large spliced-to-unspliced count ratio as direct evidence that splicing is fast. The ratio compares two abundances, not a conversion rate. The pools may have different degradation rates, and the cell may not be at a stable state. Rate inference requires additional assumptions and information; it cannot be obtained from a single abundance ratio alone.

A second diagnostic is dimensional consistency. A number reported in molecules per minute is a flux or production rate; a number reported in inverse minutes is a per-molecule rate constant; and a molecule count is an abundance. Multiplying an inverse-time constant by a pool abundance gives a flux. Confusing these units can lead to incorrect biological conclusions even when the arithmetic is correct.

Experimental strategies can provide information beyond a static snapshot. For example, pulse-labeling or pulse-chase designs track newly synthesized RNA over time and can help distinguish production and removal processes. Such measurements still require assumptions about labeling, processing, and sample comparability. A simple first-order summary may be useful, but it can obscure multiple introns, processing delays, transcript-specific stability, or cell-to-cell heterogeneity.

Comparing the Quantities