Writer, eraser and reader mechanism studies
What a modifying enzyme actually does, site by site. Perturb the enzyme, then compare the modification before and after — the difference is the enzyme’s footprint.
01The question
A mechanism study asks which positions a writer, eraser or reader acts on, and what changes at those positions when the enzyme is perturbed. The unit of the answer is a site, so the readout has to resolve one.
The design is always comparative. A single profile describes one state; the informative result is the difference between an enzyme-perturbed sample and its matched control, measured on the same positions in the same assay.
That makes pairing important. Where the perturbation changes which transcripts are present, a modification difference has to be read alongside the transcript-level data rather than on its own.
02Which route fits which enzyme question
Single-Base RNA Modification Sequencing
The route when the conclusion names positions: each site is called individually and its level compared between perturbed and control samples.
Site-levelposition & level m6Am6A Modification Seq (CAM-seq)
For m6A writers, erasers and readers acting on mRNA and lncRNA.
Site-levelposition & level m3Cm3C Modification Seq (m3C-IP-seq)
Antibody enrichment with misincorporation, for m3C enzymes.
Site-levelposition SCREENRNA Modification RIP-Sequencing
Used first where the perturbation is broad and the aim is to see which transcripts move.
Transcriptenrichment03What a mechanism result contains
- Matched site tables. The same positions called in both arms of the experiment, with their modification ratios.
- A difference table. Sites that gained or lost modification, with fold change and significance.
- Enzyme-to-site linkage. The set of positions whose level tracks the enzyme, which is the core of the mechanism argument.
- Motif context. Whether the affected positions share a sequence context, which either supports or complicates the linkage.
- Transcript-level context. Abundance data for the affected transcripts, so a change in modification is not confounded with a change in expression.
04Setting up the comparison
- Define the perturbation: knockdown, knockout, overexpression or inhibitor, and the matched control.
- Decide whether the readout needs positions or a profile. A named-enzyme argument normally needs positions.
- Run both arms in the same assay so the two site tables are directly comparable.
- Read the difference table against the perturbation, and check whether expression changed underneath the modification signal.
| What you want to conclude | Route | Request Quote |
|---|---|---|
| Which positions an enzyme acts on | Single-base sequencing on both arms | |
| Which transcripts shift with the enzyme | RIP-sequencing on both arms | |
| Whether a level change is real | LC-MS as an independent readout |
05FAQ
Why do both arms have to be run in the same assay?
Because the comparison is between modification ratios at the same positions, and a ratio is only meaningful against one measured the same way. Running the two arms on different methods introduces a difference that is not the perturbation, so the perturbed sample and its matched control are processed through one workflow.
What if knocking down the enzyme also changes transcription?
It often does, and that is why the analysis pairs the modification difference with transcript abundance, so a change in modification is separated from a change in how much transcript exists. Without that second readout, a site that loses modification and a site on a transcript that has simply become less abundant look the same.
Can a reader protein be studied the same way?
Yes, where binding can be perturbed. The readout is still the modification state of the positions whose fate the reader is expected to influence, and the argument rests on those positions changing when the reader is removed. What the modification data alone cannot show is that the reader is what caused the effect.
Do I need to know the enzyme before starting?
No. A broad screen first is a normal route: compare the profile across the perturbation, then take the transcripts that moved into site-level work. Starting that way costs nothing in interpretation as long as the screen is read as a level rather than a position, and the positions are then called in the same two arms.