What Is Recombinase Polymerase Amplification?

Recombinase polymerase amplification, usually called RPA, is a method of copying specific stretches of DNA quickly and at low temperatures, without the expensive equipment that traditional lab techniques require. It works at around 37–42°C and can detect as few as one to ten copies of a target DNA sequence in under twenty minutes.¹ That combination of speed, sensitivity, and simplicity has made RPA one of the most talked-about tools in point-of-care diagnostics, though the technology carries its own set of trade-offs that are worth understanding.

How RPA Actually Works

If you have ever heard of PCR (polymerase chain reaction), the gold-standard method for amplifying DNA in laboratories, RPA is easiest to understand in contrast. PCR uses repeated cycles of heating and cooling to pull apart double-stranded DNA so that new copies can be built. That heating-and-cooling cycle requires a thermocycler, which is a precision instrument that costs thousands of dollars. RPA sidesteps the heating step entirely by using proteins borrowed from living cells’ own DNA-repair machinery.

Three types of proteins do the heavy lifting. A recombinase protein binds to short synthetic DNA primers and guides them to the matching spot on the target DNA, physically pushing the two strands apart without heat. A single-stranded DNA-binding protein (SSB) then coats and stabilizes the strand that has been displaced, preventing the two strands from snapping back together.² Finally, a strand-displacing DNA polymerase extends the primers, reading the template and building new DNA as it goes, all while continuing to peel the strands apart ahead of itself.³ Each new double-stranded copy becomes a template for the next round, and the whole process snowballs exponentially, much the way PCR does, but at a constant gentle temperature rather than through dozens of hot-cold cycles.

Temperature, Speed, and What That Means in Practice

The standard operating window for RPA is 37–42°C, roughly human body temperature.⁴ That is a much gentler requirement than PCR’s repeated spikes to around 95°C. In practical terms, it means the reaction can be kept warm by a simple battery-powered heater, a chemical heat pack, or even body heat in a pinch. One study testing an RPA assay for HIV DNA found that amplification succeeded across a broad range of incubation temperatures, from 31°C all the way up to 43°C, with performance only dropping off above 44°C and below about 27°C.⁵ That forgiveness around temperature is a big deal for field diagnostics, where tightly controlled lab conditions are a luxury.

Speed is the other standout feature. Most RPA reactions reach detectable levels of amplified DNA within five to twenty minutes. A real-time RPA assay targeting the neurotoxin gene of the bacterium that causes tetanus, for instance, achieved detection in under twenty minutes at a limit of twenty copies per reaction.⁶ Compare that to a standard PCR run, which typically takes one to two hours of cycling time alone, plus the time spent on sample preparation and gel electrophoresis. RPA compresses the whole workflow dramatically.

How RPA Compares to PCR and Other Isothermal Methods

RPA is not the only isothermal amplification technique. LAMP (loop-mediated isothermal amplification) has been around longer and is widely used. In a head-to-head comparison of four isothermal methods for detecting chicken DNA adulteration in food products, researchers found RPA was the best fit for situations requiring multiplex detection of several targets at once, while LAMP performed better for detecting a single target at lower cost. RPA reagents cost roughly ten times more than LAMP reagents, which is one of the main reasons LAMP still dominates in some settings.⁷ LAMP also requires four to six primers per target compared to RPA’s two, making LAMP primer design considerably more complex.

Against real-time PCR, RPA can hold its own in sensitivity and sometimes outperform it. In a study comparing the two methods for detecting a liver fluke parasite (Fasciola hepatica) in human stool samples, RPA achieved 87% sensitivity versus 66% for PCR, and both were 100% specific, meaning neither gave false positives. RPA also caught 47% of infections that microscopy missed, compared to 26% for PCR.⁸ Those kinds of results explain why RPA has generated so much excitement, though it is important to note that performance varies by assay and target. Not every RPA assay outperforms its PCR equivalent.

Reading the Results

After amplification, you still need a way to see whether the reaction found what it was looking for. Several detection formats have been paired with RPA, and the choice usually depends on where the test is being used.

Lateral flow strips: These work like a pregnancy test. Amplified DNA labeled with specific tags migrates along a paper strip and produces a visible line if the target is present. One early device made from layers of paper, glass fiber, and plastic was able to amplify ten copies of HIV DNA to detectable levels in fifteen minutes using lateral flow strips as the readout, all without laboratory infrastructure.⁹
Fluorescence (Exo probes): A fluorescent probe sits in the reaction mix. When the target DNA is present and amplified, the probe is cleaved by the exonuclease activity built into certain RPA enzyme formulations, releasing a fluorescent signal that can be tracked in real time with a small portable reader.
CRISPR-based readouts: A more recent approach couples RPA’s amplification power with the precise recognition ability of CRISPR-Cas enzymes, which is discussed in more detail below.

Lateral flow strips are the cheapest and most field-friendly option, since results are visible to the naked eye. Fluorescence gives a quantitative, real-time curve, useful in settings where you want to measure how much target DNA is present, not just whether it exists. Each approach has a different trade-off between cost, complexity, and the amount of information you get back.

Pairing RPA with CRISPR for Ultrasensitive Detection

One of the most consequential developments involving RPA is its integration with CRISPR-based diagnostics. The platform known as SHERLOCK (Specific High-sensitivity Enzymatic Reporter unLOCKing) uses RPA as a pre-amplification step, then feeds the amplified material to CRISPR-Cas13 or Cas12 enzymes that recognize a specific sequence and trigger a detectable signal, either fluorescence or a color change. The entire process, from sample to result, takes under an hour with about fifteen minutes of setup.¹⁰

This pairing addresses one of RPA’s weaker points: specificity in complex samples. On its own, RPA occasionally amplifies nonspecific products because the reaction temperature is low enough that primers can bind imperfectly. Adding a CRISPR-Cas step provides a second, independent layer of sequence verification, significantly reducing false positives. A one-tube assay combining RPA with Cas12a for mpox virus detection achieved a sensitivity of one copy of viral DNA within thirty minutes.¹¹ Another group built paper-based microfluidic devices that integrated RPA with CRISPR/Cas12a to detect high-risk HPV strains linked to cervical cancer, achieving sensitivity down to one picomolar concentration in about an hour.¹²

During the COVID-19 pandemic, researchers used centrifugal microfluidic chips to run multiplex reverse-transcription RPA assays that simultaneously detected three SARS-CoV-2 genes in thirty minutes, with sensitivity reaching as low as ten RNA copies per reaction for one of the targets.¹³ These chip-based approaches are moving RPA toward fully automated sample-to-answer devices that a non-expert can operate.

Where RPA Is Being Used in the Field

The original promise of RPA was diagnostics beyond the lab, and that promise has been partially realized. For neglected tropical diseases in under-resourced regions, RPA’s ability to run at constant temperature without complex equipment makes it a strong candidate for decentralized testing. A 2025 review noted that RPA’s compatibility with portable detection platforms makes it well suited for deployment in field settings that lack traditional laboratory infrastructure.¹⁴ Applications have expanded beyond human medicine into veterinary diagnostics, food safety testing, plant pathogen detection, and environmental monitoring.¹⁵

The earliest lab-on-a-chip systems for RPA used foil-based centrifugal cartridges with pre-stored dry and liquid reagents, needing only incubation at 37°C and a fluorescence reader to deliver fully automated nucleic acid analysis.¹⁶ Since those early demonstrations, the devices have gotten smaller and cheaper. Some current prototypes are entirely disposable, designed to be used once and discarded.

Reagent Stability and the Cold-Chain Problem

For any diagnostic technology destined for tropical or rural settings, the question of reagent stability without refrigeration is critical. Standard RPA kits ship freeze-dried, which helps, but the enzymes in the pellet can lose activity over time if stored at high ambient temperatures. The evidence on this front is encouraging. In the first peer-reviewed study of long-term RPA reagent stability without a cold chain, reagents stored at room temperature (25°C) for up to twelve weeks still detected ten copies of HIV DNA. Even at 45°C, reagents held up for three weeks before sensitivity began to decline.¹⁷

Researchers have been pushing that stability further. Lyophilization with sugar-based protectants, particularly combinations of trehalose, pullulan, and mannitol, has improved enzyme preservation in dried reagents.¹⁸ A separate approach replaced one of the conventional enzymes in the RPA mix, pyruvate kinase, with a thermostable version sourced from a heat-loving bacterium. The resulting lyophilized reagent showed higher performance than the standard formulation, suggesting that building the kit from heat-resistant enzyme variants could be a path toward reagents that survive months in a backpack without refrigeration.¹⁹

Limitations Worth Knowing About

RPA has real weaknesses that often get glossed over in the excitement about its field potential. The most frequently cited is cost. The commercial kits, currently sold primarily by one company (TwistDx, now part of Abbott), are significantly more expensive per reaction than PCR or LAMP reagents. That price gap puts a ceiling on RPA’s scalability for large surveillance programs in low-income countries, which is precisely the setting where RPA’s technical advantages would shine brightest.

Primer design can also be finicky. RPA primers are typically longer than PCR primers, usually around 30 to 35 bases, and the reaction’s tolerance for mismatches is both a strength (it can detect slightly mutated targets) and a weakness (it can amplify off-target sequences). Nonspecific amplification products occasionally appear, especially in longer incubations. This is one reason the CRISPR-coupled approaches discussed earlier have become so popular: the second layer of recognition compensates for RPA’s occasional imprecision.

Sample preparation remains a bottleneck, though it is less severe than with PCR. RPA tolerates common inhibitors like humic acid, hematin, tannic acid, and melanin better than many competing methods. One study demonstrated that a crude blood sample could be processed with a simple heating step and a chemical lysis buffer, skipping RNA purification entirely, and still produce reliable results.²⁰ Another study showed that the RPA assay tolerated raw buccal swab material or a small volume of blood directly in the reaction mix.²¹ That tolerance is impressive, but “minimal preparation” is not the same as “no preparation,” and getting a clean enough sample in field conditions still takes some skill and care.

Quantification is another area where RPA lags behind real-time PCR. Because the amplification is so fast and the reaction kinetics are less uniform than PCR’s tightly controlled thermal cycles, the relationship between starting DNA quantity and signal timing is noisier. You can reliably say “yes, the target is present” or “no, it isn’t,” but pinning down exactly how much target was in the original sample is harder. Digital RPA formats, which partition the reaction into thousands of tiny droplets or chambers, are being explored to solve this, but the technology is still mostly in academic labs.

Multiplexing and Detecting Several Targets at Once

Detecting multiple pathogens in a single reaction, known as multiplexing, is possible with RPA but comes with extra complications. Each target requires its own pair of primers, and all those primers have to coexist in the same tube without interfering with one another. Researchers have demonstrated multiplex RPA using primers labeled with different chemical tags, then separating the results on a lateral flow strip with multiple test lines. In one such system, a biotin-labeled forward primer required double the concentration of the other labeled primers to produce a visible signal, and further increases actually interfered with the detection of a different label.²² This kind of cross-talk between primer sets makes multiplex RPA assay development considerably more labor-intensive than designing a single-target test.

The microfluidic approach offers an alternative route: instead of cramming all targets into one tube, you physically separate them into distinct reaction chambers on a chip, each preloaded with its own primer set. The centrifugal SARS-CoV-2 chip mentioned earlier used this strategy, running four reactions in parallel on a single disposable disc.¹³ This sidesteps the primer interference problem but adds manufacturing complexity to the device.

The Patent and Supply Landscape

A less-discussed constraint on RPA’s adoption is the intellectual property situation. The core RPA technology was developed by TwistDx in the United Kingdom, which was acquired by Abbott in 2021. For years, TwistDx held key patents covering the fundamental reaction chemistry, which meant other companies could not freely manufacture or sell competing RPA kits. This limited the number of suppliers and kept kit prices high. Several of the original patents have begun to expire or are nearing expiration, and academic groups in Asia and Europe have started publishing work on open-source enzyme combinations using recombinases and polymerases from thermophilic bacteria rather than the proprietary TwistDx components.⁴ If these alternative enzyme systems mature, they could break the single-supplier bottleneck and bring reagent costs down sharply, making RPA viable for the kind of mass-scale deployment that organizations fighting neglected tropical diseases need.

The technical quality of these alternative formulations is still catching up. Published studies show promising amplification but have not yet matched the commercial kit’s optimized consistency and shelf stability across a wide range of targets. Nonetheless, the trend toward open and diversified reagent sources is one of the most consequential developments in the RPA space, because the technology’s greatest limitation right now is not scientific performance but availability and cost.