What Is RNA Capture and When Do You Need It

Team Dynamic Matrices | 25 March 2026

RNA extraction gives you all the RNA in a sample. RNA capture gives you the RNA you asked for. That distinction sounds small and it decides whether a workflow is feasible for a lot of the samples that matter: scarce ones, degraded ones, ones where the signal is a fraction of a percent of the total. This post is a plain explanation of what capture is, how it differs from its neighbours, and when it is the right tool.

Three things people confuse

Extraction separates nucleic acid from everything else (protein, lipid, cell debris). Output: total RNA. Selectivity: none by sequence.

Depletion removes an unwanted class, usually ribosomal RNA, and keeps everything else. Output: total RNA minus rRNA. Selectivity: negative, against a known set of sequences.

Capture (also called targeted enrichment or pull-down) selects a defined set of sequences and discards everything else. Output: only the targets. Selectivity: positive, for a known set of sequences.

Depletion and capture are the same chemistry pointed in opposite directions: in one you keep the supernatant, in the other you keep the pellet. Which is right depends on whether you know what you want.

How capture works

All sequence-specific capture rests on hybridisation. A probe complementary to the target binds it; something then separates probe-target pairs from the rest. The separation step is where methods differ:

• Bead-based hybrid capture. Biotinylated probes, streptavidin beads, magnet. The standard for exome and panel enrichment (Gnirke et al., Nature Biotechnology, 2009). Capacity is bead surface; background is bead surface.
• Solution-phase capture with precipitation. Probes carry a handle that lets the hybrid be precipitated or pelleted without a solid surface.
• Phase-separation capture (LASSO). Catcher strands on a soluble polymer; hybridisation crosslinks the polymer so it condenses into a pellet; release by strand displacement under native conditions. No beads, no enzymes (Dynamic Matrices, Angewandte Chemie, 2025).
• Aptamer or antibody capture. Bait recognises structure or a bound protein rather than sequence. Useful for RNPs; not sequence-programmable.

When you need it

1. The target is a small fraction of the total. A viral transcript in a host sample, a fusion transcript in a tumour, a set of biomarkers in plasma. Sequencing everything wastes depth on background; capturing first concentrates reads where they matter. For a target at 0.1 percent of the input, capture can raise it to the majority of the library.

2. The sample is degraded. FFPE and cell-free RNA are fragmented. Capture probes tiled across a target recover fragments regardless of whether they carry a poly(A) tail or span an rRNA-depletion probe site. Cieslik et al. (Genome Research, 2015) showed capture RNA-seq outperforms depletion on archival tissue.

3. The sample is scarce. At nanogram input, every bind-wash-elute step costs a fraction of what you have. Capturing the targets in one step, with a chemistry that has no surface to lose material on, keeps more of the little there is.

4. You need the RNA with its partners intact. To study an RNA-protein complex, an RNA structure, or an enzyme bound to RNA, the release step must not denature. Native-condition capture and release is the only way to get the complex back working.

When you do not need it

If you want the whole transcriptome from abundant, intact RNA, use extraction plus depletion and sequence everything. Capture requires you to know the targets in advance, adds a step, and can introduce coverage bias across a panel if probes are not thermodynamically matched. Discovery experiments should not be captured; validation and diagnostic experiments usually should.

What to ask of a capture method

• Specificity: what fraction of captured material is on-target, measured on your sample type?
• Uniformity: across a multi-target panel, how even is per-target recovery?
• Capacity: how much target can one reaction hold before it saturates?
• Release conditions: denaturing, or native?
• Redesign cost: what does it take to add or change a target?
• Reagent stability and enzyme dependence: what does the cold chain look like?

Our pull-down method comparison and multiplexed panel post go into each of these.