Digital microfluidics methods for nucleic acid detection: A mini review

Abstract

Many serious infectious diseases have occurred throughout human history. Rapid and accurate detection as well as the isolation of infected individuals, through nucleic acid testing, are effective means of containing the spread of these viruses. However, traditional nucleic acid testing methods rely on complex machines and specialized personnel, making it difficult to achieve large-scale, high-throughput, and rapid detection. In recent years, digital microfluidics has emerged as a promising technology that integrates various fields, including electrokinetics, acoustics, optics, magnetism, and mechanics. By leveraging the advantages of these different technologies, digital microfluidic chips offer several benefits, such as high detection throughput, integration of multiple functions, low reagent consumption, and portability. This rapid and efficient testing is crucial in the timely detection and isolation of infected individuals to prevent the virus spread. Another advantage is the low reagent consumption of digital microfluidic chips. Compared to traditional methods, these chips require smaller volumes of reagents, resulting in cost savings and reduced waste. Furthermore, digital microfluidic chips are portable and can be easily integrated into point-of-care testing devices. This enables testing to be conducted in remote or resource-limited areas, where access to complex laboratory equipment may be limited. Onsite testing reduces the time and cost associated with sample transportation. In conclusion, bioassay technologies based on digital microfluidic principles have the potential to significantly improve infectious disease detection and control. By enabling rapid, high-throughput, and portable testing, these technologies enhance our ability to contain the spread of infectious diseases and effectively manage public health outbreaks.

I. INTRODUCTION

Throughout human history, many major diseases have been caused by viruses, leading to global security emergencies. In early 2020, a novel coronavirus spread globally, posing a serious threat to human life.¹ To control the spread of the epidemic, timely detection of infected individuals or carriers of the virus and their isolation through nucleic acid testing is crucial. The first step in the traditional nucleic acid testing process involves sample collection and nucleic acid purification.² Samples can include blood, urine, saliva, and other body fluids. After collection, the cells need to be lysed through pre-processing to release cell membranes, proteins, polysaccharides, DNA, RNA, and other substances. Cell membranes, proteins, and other substances may interfere with subsequent amplification steps, so nucleic acid purification is necessary. This can be achieved using methods such as magnetic beads and concentrated salt methods to remove impurities. The second step is nucleic acid amplification, with the most commonly used technique being the polymerase chain reaction (PCR).^3–7 PCR amplifies the detection signal by increasing the specific target fragment in three steps: denaturation, annealing, and extension. After amplification, nucleic acid detection is performed using fluorescence detection, colorimetric methods, or turbidity detection methods.

The traditional nucleic acid detection process is complex, time-consuming, and requires skilled professionals to perform. This poses challenges for remote and primary hospitals with limited resources and environmental equipment, as they struggle to implement the traditional method and improve detection efficiency.

Microfluidic chips, as an emerging technology in nucleic acid detection, offer numerous advantages in terms of high throughput, high sensitivity, and automation.⁸

Using digital microfluidics technology, samples and reagents ranging from microliters to nanoliters can be automatically separated and detected with high resolution and sensitivity. It has the following advantages: (1) Digital microfluidics can achieve individual droplet manipulation without the need for external devices such as channels, pumps, valves, and mixers, thereby reducing the barrier for equipment usage. (2) It allows for free manipulation of droplet movement paths without the need for complex chip structures, making it highly adaptable to different scenarios. (3) It does not cause channel clogging. (4) By introducing oil-based fluids, the chip can be sealed, and droplets can be kept independent, preventing sample contamination. With these advantages, digital microfluidics technology has been widely applied in the construction of biochemical pharmaceutical synthesis and analysis, nucleic acid detection, cell culture and processing, as well as immune diagnostics platforms or systems.

In recent years, there have been many studies using digital microfluidics technology for nucleic acid detection. For example, dPCR technology can directly obtain the copy number of target molecules, eliminating the need for standard curves or reference materials to determine the copy number of target genes, thus enabling absolute quantification of target genes.^9–11 The basic principle of dPCR was proposed by Skes et al. in 1992.¹² Although it was not yet called “digital PCR” at that time, based on this publicly published method, the basic experimental procedure of digital PCR has been established. In addition, researchers have combined digital microfluidics technology with LAMP to develop digital LAMP (dLAMP) technology,^13,14 which inherits the advantages of LAMP, such as low instrument complexity, high specificity, accuracy, and strong tolerance to inhibitors. dPCR and dLAMP¹⁵ technologies are widely used in point-of-care testing (POCT). Furthermore, in recent years, researchers have combined CRISPR technology with digital microfluidics technology to develop highly sensitive and specific nucleic acid amplification-free digital detection methods. For example, Shan et al. used negative pressure-driven microfluidic chips to generate thousands of monodisperse droplets with a size of 30 μm in just 2 min.¹⁶ By confining individual target RNA recognition events within an isolated droplet, the activated Cas13a collateral cleavage products can accumulate in a droplet. Combining microfluidic droplets with CRISPR-Cas13a, SARS-CoV-2 RNA can be easily detected within 30 min, with a detection limit of 470 aM. In conclusion, digital microfluidics technology has great development prospects in the field of nucleic acid detection.

In this review, we outline how the properties of digital microfluidics have been leveraged for nucleic acid detection, discuss different digital microfluidics strategies (which today form part of most microfluidics-based diagnostic assays), and review the optimization of digital microfluidics for POC applications, with a focus on assay readouts and sample preparation. We also provide an overview of the emerging biomedical applications of the technology and discuss open challenges and opportunities.

II. NUCLEIC ACID DETECTION

In general, purified nucleic acids tend to have low concentrations and are not easily detected directly. Therefore, amplification of the nucleic acid signal is necessary. Based on different principles of amplification, detection can be categorized into three groups, as illustrated in Fig. 1: the first group involves amplification of the target fragment, wherein the detection is achieved by base pairing, resulting in an increase in the target fragment; the second group is based on CRISPR technology, which recognizes the target fragment and triggers protein activation in response to the reporter molecule; and the third group is nucleic acid detection based on local surface plasmon vibration technology.

FIG. 1.

Three ways to test for nucleic acids.

A. Amplification

Currently, amplification is the most commonly used method for nucleic acid detection. Based on the temperature requirements, these methods can be divided into two categories. The first category is polymerase chain reaction (PCR), which is considered the gold standard for nucleic acid detection. In the PCR system, polymerase, primers, and templates are key components. The process involves denaturation of the double-stranded DNA at approximately 95 °C, followed by primer binding to the template at 60 °C, and extension of the primers along the template at 72 °C through the action of polymerase, leading to amplification. However, PCR reactions are dependent on temperature changes and are not suitable for point-of-care testing (POCT) assays. As a result, an alternative method emerged, known as isothermal amplification. This includes techniques like loop-mediated isothermal amplification (LAMP),^17,18 recombinase polymerase amplification (RPA),¹⁹ and helicase-dependent amplification (HDA).^20,21 LAMP requires a temperature of 65 °C and typically involves the use of 4–6 primer pairs. It offers good specificity and holds great potential for POCT when combined with microfluidic technology. RPA, on the other hand, can be performed at a temperature range of 37–42 °C and can utilize the surface temperature of the human body as a heat source, making it suitable for wearable devices.

B. CRISPR-based methods for detection

Since its discovery in 1980, the CRISPR system has found widespread use in gene editing and other applications.²² CRISPR/Cas systems can be classified into two categories based on the nature of the ribonucleoprotein effector complexes. Class I systems consist of complexes composed of multiple effector proteins, while Class II systems involve a single crRNA-binding protein that can be used for nucleic acid detection. Cas9, guided by both crRNA and tra-crRNA, can specifically cleave DNA sequences, relying on pre-amplification (1). Cas12, guided solely by crRNA, can also specifically cleave DNA sequences and has non-specific DNA cleavage activity. Likewise, Cas13 can specifically cleave RNA sequences with crRNA guidance and exhibit non-specific RNA cleavage activity. Both Cas12 and Cas13 systems have been applied to nucleic acid detection in various ways. They can be combined with pre-amplification techniques,²³ such as DETECTR,²⁴ SHERLOCK,²⁵ CARMEN,²⁶ and others. Additionally, these systems can be used for direct amplification-free nucleic acid diagnosis.^27,28 Fozouni et al.²⁹ developed a device that enables the direct detection of SARS-CoV-2 using cell phones, without the need for amplification.

C. Surface plasmon resonance for detection

Surface plasmon vibration technology is an optical detection method used for real-time monitoring of biomolecular interactions.³⁰ Typically, an oligonucleotide probe is attached to the sensor surface, and if the sample being tested contains the target fragment, it will bind to the surface probe, resulting in a change in the refractive index of the sensor surface. By establishing a correlation between nucleic acid concentration and refractive index, quantitative measurements can be obtained. This assay does not rely on traditional fluorescence detection equipment and requires only a small sample size. However, the application of this method for detecting nucleic acids in complex samples still poses challenges.

III. DIGITAL MICROFLUIDICS

Nucleic acid detection can be accomplished through various methods such as amplification, CRISPR-based technology, or SPR technology. The above methods are more difficult to realize the detection of single molecules. Digital microfluidics, on the other hand, can separate single molecules by dilution so that each reaction chamber contains 0 or ≥1 copy of the target DNA. A Poisson correction can be added to the results to account for reaction chambers containing multiple molecules and to estimate the absolute number of target sequences.^31–35

A. Statistical analysis

The theory of digital microfluidics is based on the theory of Poisson distribution in statistics.³⁶ The probability function of the Poisson distribution can be expressed as follows:

$$P(X=k)={\displaystyle \frac{{\lambda }^k}{k!}{e}^{-\lambda },k=0,1,\dots ,}$$ (1)

where λ is the average number of random events per unit time (or unit area). When the model is applied to describe the distribution of nucleic acid molecules, λ is the ratio of the number of positive molecules to the total number of molecules,

$$\lambda ={\displaystyle \frac{b}{n},}$$ (2)

where b represents the number of positive molecules and n represents the number of the total molecules. When k equals 0, the meaning of P(k = 0) is the probability of negative molecules, which can be deduced from the following relationship:

$$P(X=0)=1-{\displaystyle \frac{b}{n}=e^{-\frac{b}{n}}.}$$ (3)

Then, the relationship between the concentration c and the volume of each chamber v can then be established as follows:

$$c={\displaystyle \frac{b}{nv}.}$$ (4)

The above equations can be combined to calculate the concentration formula as follows:

$$c={\displaystyle \frac{-\ln \left(1-{\displaystyle \frac{b}{n}}\right)}{v}.}$$ (5)

In summary, when quantifying nucleic acids using digital microfluidics, the concentration of nucleic acids in the corresponding sample can be calculated simply by knowing the volume of each test chamber and the proportion of chambers with positive molecules to the total chambers.

The theoretical analysis above was validated using specific data. The same microstructure was divided into 56, 400, 5000, and 20 000 pores, and the same number of target molecules were injected into the microstructure.³⁷ As more holes are added, the size of the holes is gradually partitioned into smaller volumes as illustrated in Fig. 2(a). Calculating the Poisson distribution for each of the four types separately, it is found that 99.5% of the wells are digitized as 0 or 1 as the number of holes increases as illustrated in Fig. 2(b).

FIG. 2.

Data processing. (a) Micro-structures with 56, 400, and 20 000 wells. As the number of well increases, the size of well gets partitioned into smaller volumes. (b) Poisson probability distribution when injecting 2000 copy genes in 56 (A), 400 (B), 5000 (C), and 20 000 wells (D). Reproduced with permission from Lee et al., Genom. Inform. 19(3), e34 (2021). Copyright 2021 Author(s), licensed under a Creative Commons Attribution (CC BY) license.³⁸

B. Digital microfluidic methods for nucleic acid detection

The increased specific surface area of microdroplets reduces reaction time and improves the performance of biochemical performance of analytical assays, reducing sample and reagent consumption. In contrast to conventional macroscopic. Compared with conventional macroscopic systems, microfluidics manipulates droplet systems at the micro- and nano-levels and has unique properties compared with ordinary droplets.^39,40 (1) Microdroplets can be considered as a large number of small-sized isolated units, which are very suitable for single-cell or single-molecule analysis.⁴¹ (2) Rapid mixing and negligible thermal inertia make microdroplets very suitable for single-cell or single-molecule analysis. (3) The two immiscible liquids and the interface between them offer the possibility of new reactions. For digital microfluidics, the most important thing is to generate stable and homogeneous droplets. However, at the same time, the volume of each chamber needs to be satisfied to be small enough when using this formula to calculate the concentration, which is currently a difficulty in the field of digital microfluidics. Here, we classify the currently available digital microfluidic chips for nucleic acid detection into the following three types based on the microdroplet travel principle.

1. Chip-in-a-tube

Using benchtop centrifuges, highly efficient and productive production of monodisperse emulsion droplets is achieved. The process involves centrifugal rotation, where the continuous aqueous phase is dispersed into monodisperse droplet jets in the air through a microchannel array (MiCA). These droplets are then immersed in oil, resulting in the formation of a stable emulsion,⁴² as depicted in Fig. 3. This method ensures effective droplet dispersion while minimizing the risk of cross-contamination. The MiCA, which is a disk with a hydrophobic monolayer beneath it, plays a crucial role. As the water phase is propelled toward the bottom of the tube by centrifugal force, it continuously undergoes pinching at the MiCA nozzle, aligning itself as water droplets that subsequently enter the receiving oil, aligning further as an anhydrous emulsion. Droplets of various sizes can be generated by adjusting the x-rotation speed, the number of channels, or the size of the MiCA. The notable advantage of this approach lies in the core design of MiCA, which can be easily mass-produced using existing technology and is highly reproducible. Additionally, this method utilizes a benchtop centrifuge, a widely available instrument in biological and chemical laboratories. Its seamless integration allows for the efficient generation of emulsions in a highly parallel manner with excellent scalability. Moreover, the fabrication of MiCA plates is cost-effective compared to PDMS and glass-based microfluidic chips, thanks to the high throughput capabilities of microchannel plate technology. In conclusion, this approach offers a low-cost and controllable digital microfluidic method that can be readily scaled up for various applications.

FIG. 3.

Construction and operation principle of the MiCA-emulsifier. (a) Assembly of a container with the MiCA. The main body was made of PEEK with a PTFE gasket ring. (b) The components. (c) The swing buckets with microcentrifuge tubes and MiCA inserts will flip centripetally when spinning. (d) During spinning, the centrifugal force is perpendicular to the MiCA plate, breaking the solution into small droplets, which then form emulsion in the receiving oil. (e) The emulsion stably sits at the bottom of a microcentrifuge tube after centrifugation. (f) Microphotograph of emulsion droplets after 40 thermal cycles of PCR. (g) Fluorescence microphotograph indicating the digital amplification within the emulsion. Scale bars: 100 μm. Reproduced with permission from Chen et al., Lab Chip 17(2), 235–240 (2017). Copyright 2017 The Royal Society of Chemistry.⁴²

2. Microstructure

This method involves the fabrication of numerous uniform microchambers on a substrate, with soft photolithography or etching techniques commonly employed in current research. PDMS is the most frequently used material, as it enables the processing of tiny chambers through photolithography. Once a large number of microchambers have been created, as depicted in Fig. 4(a), the key next step is to disperse the solution into each individual chamber while preventing cross-contamination. One approach to achieve this is oil-phase cutting separation, as illustrated in Fig. 4(b), which relies on the immiscibility between the oil and water phases. However, it is important to consider the pressure applied when introducing the oil phase using this method. The second method is based on mobile microfluidics, which typically involves two sliding plates. In this approach, a mixture of liquid components is applied, and the droplets are mixed by sliding the plates to achieve cross-dispersion-free droplets,^43–46 as demonstrated in Fig. 4(c). The third method relies on the hydrophilic and hydrophobic properties of the substrate,⁴⁷ as depicted in Fig. 4(d). By modifying the hydrophilic array on a hydrophobic substrate, droplets can adhere to the hydrophilic array. Additionally, the size of the microchamber can be adjusted by changing the size of the hydrophilic array. Using microstructures for digital microfluidic droplet dispersion offers several advantages. First, it allows for precise control over the size of the chamber. Second, the loss of the mixture is minimized, resulting in more accurate nucleic acid quantification. Third, it helps prevent aerosol contamination. Finally, integrating most procedures on the same device enhances the portability of the chamber-based microfluidic chip. For example, by incorporating a heater and a signal reader, this digital microfluidics approach can be integrated into a portable machine, further enhancing its portability.

FIG. 4.

Microstructure microfluidic chips. (a) Photolithography production process; (b) oil-phase cutting separation chips; (c) slip chip; (d) hydrophilic and hydrophobic chip.

In addition to microstructures that can be realized using PDMS materials, PCTE can also be used, which is a commercially available membrane with numerous pores on its surface containing a high density of uniform micro-/nanopores. Each pore serves as an individual nano-reactor for DNA amplification⁴⁸ as depicted in Fig. 5. One significant advantage of this method is the cost-effectiveness of the disposable PCTE membrane, which is priced at less than $0.10 per piece. To our knowledge, this is the most affordable way to perform digital LAMP. This membrane system opens up possibilities for point-of-care users and general laboratories by allowing them to conduct digital quantification, single-cell analysis, and other bioassays in an inexpensive, flexible, and streamlined manner.

FIG. 5.

PCTE. (a) PCTE membrane dispersion sample liquid process. (b) Mechanism for excess sample removal.

3. Dielectric wetting

In this technique, the movement, merging, splitting, and generation of microdroplets with volumes ranging from picoliters to microliters can be achieved by applying a voltage to the droplets on a hydrophobic surface and altering the contact angle between the droplets and the surface.^49,50 Furthermore, the addition of nano-particles to the droplet causes a favorable recovery of the electro-spreading characteristics of a soft surface by realizing an alteration in the effective dielectric constant of the interfacial region. This technology may open up new vistas in droplet-based microfluidics.^51,52 The digital microfluidic method based on the dielectric wetting principle shares similar advantages with traditional flow channel microfluidics, including low reagent consumption, efficient heat transfer, and easy integration. In addition to these benefits, this method presents several unique advantages: (a) Digital microfluidics can achieve individual droplet manipulation without external devices such as channels, pumps, valves, and mixers, which reduces the threshold of using devices. (b) Free manipulation of droplet movement paths can be achieved without the need for complex chip structures, which is highly applicable to different scenarios. (c) The open chip structure allows solid samples to be processed without blocking channels. (d) The chip can be closed by adding oil-based liquid, and the droplets are independent of each other to prevent contamination between samples. With the above advantages, this technology has been widely used in recent years to build platforms or systems for biochemical and pharmaceutical synthesis and analysis, cell culture and processing, and immune diagnosis.⁵³

C. Result analysis

In nucleic acid detection, the application of digital microfluidics involves dispersing nucleic acid molecules from a sample into a high-density array of microchamber reactions. Positive chambers containing target molecules exhibit amplification or the CRISPR system, while negative chambers do not contain target molecules. By quantifying the number of positive and negative chambers and applying a Poisson distribution for data correction, the concentration of the target sample can be accurately determined. This method of digital microfluidics offers several advantages: it does not rely on internal reference genes, enables absolute quantification of DNA or RNA molecules without the need for a standard curve, allows for detection of nucleic acid molecules even at a single-copy level, and exhibits increased tolerance to variations in the external environment. Based on the statistical analysis discussed earlier, it is evident that this approach facilitates precise concentration determination.

According to the aforementioned that the digital microfluidic method can achieve the dispersion of droplets, then the result analysis needs to be done. From Sec. III B, the nucleic acid concentration of the sample to be tested can be calculated by the following equation:

$$c={\displaystyle \frac{-\ln \left(1-{\displaystyle \frac{b}{n}}\right)}{v}.}$$ (6)

Taking log on both sides, we get the following formula:

$$\lg (v)=\lg \left(-\ln \left(1-{\displaystyle \frac{b}{n}}\right)\right)-\lg c.$$ (7)

Therefore, the intercept is used to calculate the concentration.^54–56 For example, the DNA concentration is calculated using this equation (Fig. 6).

FIG. 6.

Fluorescence image results of different concentrations of λDNA by using the fully integrated NA detection platform. (a) Fluorescence images of digital LAMP with serially diluted λDNA (concentrations ranging from 1/2 to 1/10000 of the stock sample concentration [1 × 10⁴ copies per μl]). (b) A regression curve obtained by plotting the fraction of positive droplets against the expected λDNA concentration (CPD) according to the Poisson distribution. (c) Comparison of measured λDNA concentrations to the expected concentrations. The measured λDNA concentrations match well with the expected λDNA concentrations according to Poisson statistics. Reproduced with permission from Mao et al., Analyst 146(22), 6960–6969 (2021). Copyright 2021 The Royal Society of Chemistry.⁴⁷

IV. CONCLUSIONS AND PROSPECTS

In recent years, digital microfluidics has gained widespread application in nucleic acid detection, including dPCR,^57–62 among others. Digital microfluidics not only complements traditional nucleic acid amplification methods but also integrates with CRISPR systems. Tian et al.⁶³ introduced a Cas13a assay inspired by the confinement effect for single-molecule RNA diagnostics, eliminating the requirement for NAA and RT. Shinoda et al.⁶⁴ have developed a platform called “CRISPR-based amplification-free digital RNA detection (SATORI)” that combines CRISPR-Cas13-based RNA detection with microchamber-array technology. Digital microfluidics offers several advantages over traditional molecular diagnostic methods, as outlined below.

Digital microfluidic technology is highly integrable and can integrate multi-functions on a chip. Moreover, digital microfluidic technology has the potential of automation, which can be widely used in the field of immediate detection. Besides, digital microfluidic technology can realize highly sensitive detection, which is conducive to early diagnosis.

Due to these advantages, digital microfluidics has been widely used in the development of platforms or systems for biochemical and pharmaceutical synthesis and analysis, cell culture and processing, as well as immune diagnosis in recent years.

However, despite the potential of digital microfluidics for nucleic acid detection, its application in the clinical setting is still limited to the laboratory stage. Currently, the Quant Studio 3D is the only commercially available instrument, highlighting the need for further advancements and improvements in this field.

First, there has been significant progress in increasing the number of microchambers that can be integrated on a chip. However, relying solely on human observation for computational analysis is insufficient to meet the growing demand. Therefore, the use of application software is necessary to perform image processing and extract quantitative information. Currently, image processing techniques employ image segmentation, which involves using the fluorescence intensity histogram to simulate the distribution of signal and background and determine the threshold value. While these algorithms can successfully separate the background into distinct groups for fluorescence images with strong and uniform signal intensity, they prove ineffective for images with poor uniformity or influenced by background noise. To address this limitation, the integration of cutting-edge algorithms such as neural networks can enhance the image processing capabilities and improve accuracy. Second, the current signal reading method for digital microfluidics primarily relies on fluorescence detection, which hampers the miniaturization of detection equipment. Thus, future directions in digital microfluidics should explore more portable signal-reading methods. One potential approach is to utilize color changes for result interpretation.^65–67 Furthermore, for home or field use, the design should incorporate simple sample preparation solutions combined with robust assays to ensure reliable results in variable or challenging situations. These scenarios may involve long storage periods, limited user training, or harsh environmental conditions. Ongoing research aims to incorporate SARS-CoV-2 sensors into masks, enabling real-time monitoring and allowing for continuous tracking and early diagnosis through direct sensing on surfaces. As individuals become more aware of personal health testing, digital microfluidic devices must move toward miniaturization and wearability.^68–72

AUTHOR DECLARATIONS

Conflict of Interest

The authors have no conflicts to disclose.

Author Contributions

Youqiang Xing: Conceptualization (lead); Funding acquisition (lead); Writing – original draft (lead); Writing – review & editing (equal). Yan Wang: Conceptualization (equal); Investigation (equal); Writing – original draft (equal); Writing – review & editing (lead). Xiang Li: Investigation (equal); Resources (equal); Writing – original draft (equal). Shangran Pang: Resources (supporting); Writing – original draft (supporting).

DATA AVAILABILITY

The data that support the findings of this study are available within the article.