SLAM-seq (thiol(SH)-linked alkylation for the metabolic sequencing of RNA) is a sequencing-based method that lets researchers distinguish freshly made RNA molecules from older ones inside living cells, revealing the dynamics of gene expression that standard RNA sequencing misses entirely. Developed and first described in 2017, SLAM-seq works by feeding cells a modified nucleoside that gets built into new RNA transcripts, then chemically converting that label so it shows up as a specific mutation pattern during sequencing. The result is a snapshot not just of which genes are active, but of how quickly each gene’s messages are being produced and broken down. That kind of temporal information has made SLAM-seq a workhorse for studying everything from how cancer drugs reshape transcription to how viruses hijack host gene expression.
Why Standard RNA Sequencing Falls Short
Conventional RNA sequencing measures everything present in a cell at the moment you extract the RNA. It gives you a parts list, but no timestamps. A gene might look unchanged between two conditions because new transcripts are being made just as fast as old ones are being destroyed, or because both production and decay shifted in lockstep. Standard sequencing cannot tell the difference. It captures the lake’s water level but not whether the river feeding it sped up or the drain at the bottom opened wider.
SLAM-seq solves this by tagging new transcripts at the moment they are made. Researchers can then computationally separate old RNA from new RNA in the same sequencing library, without needing to physically isolate labeled molecules beforehand. That separation is what opens the door to measuring transcription rates, RNA processing speeds, and decay kinetics on a genome-wide scale.
The Chemical Trick Behind SLAM-Seq
The method relies on 4-thiouridine (s4U), a sulfur-containing analog of the natural nucleoside uridine. When you add s4U to cell culture medium, cells take it up and their RNA polymerases incorporate it into freshly synthesized RNA in place of normal uridine. At this stage, the s4U-containing RNA would look nearly identical to unlabeled RNA during sequencing, so SLAM-seq introduces a chemical conversion step after RNA extraction.
The thiol group on s4U is reacted with iodoacetamide (IAA), an alkylating agent that modifies the sulfur-containing residue. Under optimized conditions, this alkylation reaches at least 98% efficiency within 15 minutes and increases the rate of thymine-to-cytosine (T-to-C) misreads during reverse transcription by about 8.5-fold, pushing the conversion rate above 0.94. In other words, almost every s4U that was incorporated into new RNA will show up as a T-to-C change in the sequencing data. Older RNA molecules, which never contained s4U, carry no such conversions.
1Nature Methods. Thiol-linked alkylation of RNA to assess expression dynamicsThe mechanism behind those T-to-C changes is straightforward. When reverse transcriptase encounters an alkylated s4U residue, it misincorporates guanine instead of the adenine it would normally pair with thymine. That G ends up in the cDNA, and in the final sequencing reads it appears as a C where you would expect a T.
2PubMed Central. Quantification of experimentally induced nucleotide conversions in high-throughput sequencing datasetsBecause the whole approach piggybacks on standard sequencing library preparation, there is no need for specialized equipment or affinity purification of labeled RNA. You feed cells s4U, extract total RNA, treat it with iodoacetamide, and prepare a normal sequencing library. The labeling information is encoded directly in the sequence reads.
Turning Raw Reads into Biological Insight
Reading T-to-C conversions out of sequencing data sounds simple in principle, but in practice the signal competes with noise. Naturally occurring single-nucleotide polymorphisms (SNPs) can look identical to s4U-induced conversions. Sequencing errors add more false positives. And as more conversions accumulate in a read, the sequence diverges further from the reference genome, which can cause mapping software to discard it or assign it to the wrong location.
A dedicated computational pipeline called SLAM-DUNK was built to handle these problems. It maintains steady mapping rates regardless of how many nucleotide conversions a read contains, recovers reads that map to multiple genomic locations, and masks known SNPs so that genuine polymorphisms are not mistaken for labeling-induced conversions.
2PubMed Central. Quantification of experimentally induced nucleotide conversions in high-throughput sequencing datasetsA complementary tool called GRAND-SLAM takes the analysis further by estimating, for each gene, the proportion of RNA molecules that are newly synthesized versus pre-existing. That proportion, combined with the known labeling duration, can be converted into an RNA half-life estimate for every gene in the dataset. The accuracy of those half-life estimates depends heavily on how long cells were exposed to s4U. Short labeling windows resolve fast-turnover transcripts precisely but produce large uncertainties for long-lived RNAs. Longer labeling periods do the opposite.
3Bioinformatics. Dissecting newly transcribed and old RNA using GRAND-SLAMChoosing the right labeling duration is therefore an experimental design decision that shapes which slice of RNA dynamics you can resolve. Researchers studying rapid transcriptional responses to a drug might label for 30 to 60 minutes. Those interested in slow-turnover structural RNAs would label for hours. Independent comparisons of computational workflows like pulseR and GRAND-SLAM across more than 11,600 human genes have found that the statistical framework matters too, with different tools yielding somewhat different confidence intervals for decay rate estimates.
4Briefings in Bioinformatics. A comparison of metabolic labeling and statistical methods to infer genome-wide dynamics of RNA turnoverDissecting Drug Responses in Cancer
One of the most impactful uses of SLAM-seq has been in cancer biology, where it helps researchers distinguish direct transcriptional targets of a drug from downstream secondary effects. Standard differential gene expression experiments often reveal hundreds or thousands of changed genes after drug treatment, but most of those changes are indirect consequences that cascade through regulatory networks. By labeling RNA only during a short window after drug exposure, SLAM-seq isolates the genes whose transcription rate changed immediately, pointing to direct regulatory targets.
The original demonstration of this approach combined SLAM-seq with pharmacological inhibition of BRD4, a protein involved in reading epigenetic marks, and MYC, a master transcription factor in many cancers. By measuring newly synthesized mRNA rather than total RNA levels, the study defined the direct gene-regulatory axis connecting these two proteins and clarified which genes truly depend on them for active transcription.
5Science. SLAM-seq defines direct gene-regulatory functions of the BRD4-MYC axisA more recent application took a similar strategy to benchmark drugs reported to inhibit MYB, a transcription factor implicated in acute myeloid leukemia. By first degrading MYB rapidly in engineered cells and performing SLAM-seq, the researchers identified 450 genes directly regulated by MYB, with 319 downregulated and 131 upregulated after MYB loss. They then tested six reported MYB inhibitors head-to-head and found stark differences: the number of genes affected varied from just 19 for one compound to over 1,100 for another, and the overlap with the true MYB transcriptional program ranged from only 1% to 34% of those 450 direct targets.
6PubMed Central. Rapid-kinetics degron benchmarking reveals off-target activities and mixed agonism-antagonism of MYB inhibitorsThose numbers are a reality check for drug discovery. A compound that affects far more genes than the target protein itself controls is clearly hitting other pathways, and one that barely overlaps with the genuine target program may not be doing what researchers thought. SLAM-seq turns that kind of evaluation from educated guesswork into quantitative measurement.
Studying Gene Regulation During Chromatin Perturbation
Beyond small-molecule drugs, SLAM-seq has been applied to study the transcriptional consequences of disrupting chromatin-associated proteins. In one study examining the histone modification H3K4me3, which marks gene promoters and is associated with active transcription, researchers used SLAM-seq with a 60-minute labeling window at 100 micromolar s4U to measure how mRNA synthesis rates changed when key regulatory proteins were rapidly depleted. Short-term depletion led to significant reductions in mRNA synthesis, revealing that the chromatin mark is not merely a passive indicator of activity but plays an active role in maintaining transcription through regulation of RNA polymerase behavior at promoters.
7Nature. H3K4me3 regulates RNA polymerase II promoter-proximal pause-releaseSLAM-seq has also been used in organisms beyond mammals. In African trypanosomes, parasites that cause sleeping sickness, RNA regulation works quite differently from mammalian cells because these organisms rely heavily on post-transcriptional control rather than transcription-factor-driven gene regulation. SLAM-seq allowed researchers to separately assess RNA processing rates and RNA half-lives on a genome-wide scale in these parasites, providing insights into how stability and processing independently shape the trypanosome transcriptome.
8Nucleic Acids Research. SLAM-seq reveals independent contributions of RNA processing and stability to gene expression in African trypanosomesTracking Host Responses to Viral Infection
Viral infections overhaul a cell’s transcriptional landscape within hours, and SLAM-seq is well suited to capture that upheaval in real time. Rather than comparing infected and uninfected cells at a single time point, researchers can track which host genes ramp up or shut down at different stages of infection by collecting SLAM-seq data at multiple labeling windows. One application of this approach examined the response of BHK21 cells to human coronavirus OC43, separately measuring steady-state and newly synthesized RNA levels to disentangle which transcriptional changes were actively driven by the virus from those that were merely residual accumulation of pre-existing transcripts.
9PubMed. Transcriptome dynamics of the BHK21 cell line in response to human coronavirus OC43 infectionThis kind of time-resolved view is especially valuable for understanding how viruses manipulate host gene expression to their advantage. Many viruses actively suppress host mRNA production while ramping up translation of their own transcripts. SLAM-seq can catch that suppression as it happens, rather than inferring it from the steady-state leftovers.
Going Single-Cell with scSLAM-Seq
A natural extension of the technology was to combine it with single-cell RNA sequencing, creating scSLAM-seq. Where bulk SLAM-seq averages across millions of cells, the single-cell version reveals how individual cells within a population differ in their transcriptional activity. The results from early scSLAM-seq experiments challenged some assumptions about what drives cell-to-cell variation. Rather than finding that different cells have fundamentally different transcriptional programs, the data pointed to gene-specific features as the primary source of heterogeneity. Promoter-intrinsic characteristics, including TBP-TATA-box interactions and DNA methylation patterns, correlated with transcriptional burst kinetics and on-off switching behavior at individual genes.
10Nature. scSLAM-seq reveals core features of transcription dynamics in single cellsIn practical terms, this means that much of the noise people see in single-cell RNA sequencing data is not random technical artifact or evidence of distinct cell states. It reflects real bursting behavior, where genes flip between active and silent states, with the frequency and duration of those bursts governed by each gene’s own regulatory architecture. scSLAM-seq can distinguish a gene that is constantly transcribed at a low level from one that fires in intense bursts with silent intervals between them, even if both produce similar average mRNA counts.
The Toxicity Trade-Off
The modified nucleoside s4U is not entirely inert. While lower concentrations are generally well tolerated, concentrations above 50 micromolar, which are the typical range used in mRNA labeling experiments, can interfere with the production and processing of ribosomal RNA. This inhibition of rRNA synthesis triggers a nucleolar stress response: nucleophosmin relocates from the nucleolus to the nucleoplasm, the tumor suppressor p53 is induced, and cell proliferation slows down.
11RNA Biology. 4-Thiouridine inhibits rRNA synthesis and causes a nucleolar stress responseThis is more than a technical footnote. If the labeling compound itself is changing cell behavior, researchers risk measuring transcriptional responses to s4U rather than to the experimental condition they actually care about. The standard workaround involves using the lowest effective s4U concentration, keeping labeling periods short, and always including s4U-treated controls alongside the experimental conditions. Some groups have also benchmarked different s4U concentrations in their specific cell type to find a sweet spot where labeling is robust enough for detection but mild enough to avoid triggering the stress pathway. The fact that many published studies use 100 micromolar s4U for one hour suggests that in many cell lines, this concentration is tolerable over short periods, but the potential for artifacts should always be evaluated.
How SLAM-Seq Compares to Similar Methods
SLAM-seq is not the only metabolic labeling approach that converts nucleoside analogs into detectable mutations during sequencing. TimeLapse-seq, developed around the same time, uses the same s4U labeling step but converts the thiol group using a different chemical reaction involving 2,2,2-trifluoroethylamine combined with an oxidizing agent rather than iodoacetamide.
12Nature Communications. Benchmarking metabolic RNA labeling techniques for high-throughput single-cell RNA sequencingBoth approaches share the same core principle and produce the same type of T-to-C conversion signal, but they differ in the specifics of their chemistry and efficiency. Head-to-head comparisons have been carried out to evaluate how each performs in high-throughput single-cell settings, where sensitivity and labeling efficiency at the single-molecule level become critical.
A distinct flavor of the technology called TUC-seq DUAL pushes the concept further by adding a second modified nucleoside, 6-thioguanosine (6sG), alongside s4U. Because the conversion chemistry for 6sG is compatible with the iodoacetamide treatment used for s4U, both labels can be applied in sequential pulses and detected simultaneously in the same sequencing library. This dual-labeling setup enables researchers to track two successive waves of transcription within the same experiment, providing particularly precise measurements of RNA decay by comparing how the first-wave and second-wave transcripts evolve over time.
13PubMed Central. Thioguanosine Conversion Enables mRNA-Lifetime Evaluation by RNA Sequencing Using Double Metabolic Labeling (TUC-seq DUAL)Another variant, SLAM-ITseq, takes the method in vivo. By expressing the enzyme uracil phosphoribosyltransferase in specific cell types, researchers can restrict s4U incorporation to those cells even in a complex tissue. This means SLAM-seq can be applied to study cell-type-specific transcriptional dynamics in living organisms rather than only in cell culture, though the experimental setup is considerably more involved.
2PubMed Central. Quantification of experimentally induced nucleotide conversions in high-throughput sequencing datasetsWhere the Method Is Headed
The trajectory of SLAM-seq mirrors a broader trend in genomics: moving from static snapshots to dynamic measurements, and from bulk populations to single cells. The chemistry is now well established and the computational tools are mature enough that the technique is no longer restricted to methods-development labs. It has become a standard functional genomics assay used alongside chromatin profiling and proteomics in multimodal studies of gene regulation.
Integration with rapid protein degradation systems, like the auxin-inducible degron and dTAG technologies referenced in the chromatin and drug-benchmarking studies above, has been especially productive. These systems let you eliminate a protein of interest within minutes, then use SLAM-seq to watch the transcriptional consequences unfold in real time. The combination effectively turns SLAM-seq into a tool for mapping transcription-factor-to-gene wiring diagrams with a speed and directness that older approaches could not match. As single-cell versions continue to improve in sensitivity and cost, the same logic will extend to mapping those wiring diagrams in individual cells within tissues, capturing not just what genes respond to a perturbation but how differently each cell in a population responds.