Recent developments and future directions in point-of-care next-generation CRISPR-based rapid diagnosis

Abstract

The demand for sensitive, rapid, and affordable diagnostic techniques has surged, particularly following the COVID-19 pandemic, driving the development of CRISPR-based diagnostic tools that utilize Cas effector proteins (such as Cas9, Cas12, and Cas13) as viable alternatives to traditional nucleic acid-based detection methods. These CRISPR systems, often integrated with biosensing and amplification technologies, provide precise, rapid, and portable diagnostics, making on-site testing without the need for extensive infrastructure feasible, especially in underserved or rural areas. In contrast, traditional diagnostic methods, while still essential, are often limited by the need for costly equipment and skilled operators, restricting their accessibility. As a result, developing accessible, user-friendly solutions for at-home, field, and laboratory diagnostics has become a key focus in CRISPR diagnostic innovations. This review examines the current state of CRISPR-based diagnostics and their potential applications across a wide range of diseases, including cancers (e.g., colorectal and breast cancer), genetic disorders (e.g., sickle cell disease), and infectious diseases (e.g., tuberculosis, malaria, Zika virus, and human papillomavirus). Additionally, the integration of machine learning (ML) and artificial intelligence (AI) to enhance the accuracy, scalability, and efficiency of CRISPR diagnostics is discussed, alongside the challenges of incorporating CRISPR technologies into point-of-care settings. The review also explores the potential for these cutting-edge tools to revolutionize disease diagnosis and personalized treatment in the future, while identifying the challenges and future directions necessary to address existing gaps in CRISPR-based diagnostic research.

Keywords: Artificial intelligence; Biosensing technologies; CRISPR-Cas systems; Disease detection; Machine learning; Microfluidic platforms; Nucleic acid diagnostics.

Fig. 1

Combining AI algorithms with CRISPR-Cas diagnostics to improve point-of-care (POC) testing. The method involves several steps: nucleic acid extraction, guide RNA (gRNA) targeting disease-specific sequences, CRISPR-Cas detection, and patient sample collection. A signal indicating the presence of biomarkers is generated upon recognition of target sequences. AI algorithms enhance gRNA design, ensuring high specificity and minimal off-target effects. AI models also use data from CRISPR diagnostics to refine prediction algorithms and optimize CRISPR performance. This integrated approach enhances the accuracy and timeliness of POC tests by combining AI’s predictive power with CRISPR’s precision

Fig. 2

The graphic illustrates how machine learning algorithms and CRISPR technology can be combined to advance genetic research. It shows how CRISPR experiments generate genetic alteration datasets, which are then processed and analyzed using various machine learning models such as decision trees and neural networks. These models predict outcomes like gene knockout efficiency and off-target effects, with subsequent CRISPR experiments being guided by these predictions in an iterative feedback loop. The figure highlights how machine learning can enhance and accelerate CRISPR-based research and therapeutic development, with applications in gene target identification, CRISPR design optimization, and drug discovery support

Fig. 3

This figure illustrates CRISPR-Cas technologies for nucleic acid detection using SHERLOCKv2, SHERLOCK, and DETECTR assays. The Cas nuclease is inactive in the absence of its nucleic acid target. When the nuclease binds to its guide crRNA and recognizes a similar target (RNA for Cas13a or ssDNA/dsDNA for Cas12a), it becomes activated and cleaves off-target nucleic acids. The collateral nuclease activity is enhanced through the use of reporter probes, which consist of a fluorophore attached to a quencher via a short oligonucleotide. The figure is divided into three parts: Left: SHERLOCKv2, which allows direct detection of viral infections in body fluids using RNA or DNA as input. Middle: SHERLOCK, which amplifies nucleic acids from clinical samples using recombinase polymerase amplification (RPA). Right: DETECTR, which employs a similar nucleic acid detection method, starting with RT-RPA or RPA, followed by detection with T7 RNA polymerase, Cas13, a target-specific crRNA, and an RNA reporter that fluoresces upon cleavage

Fig. 4

CRISPR and AI-based integrated point-of-care system for rapid malaria diagnosis. This figure illustrates an integrated point-of-care (POC) system for the rapid diagnosis of malaria. AI algorithms enhance the accuracy of the system. First, nucleic acids are extracted from patient samples, after which the CRISPR-Cas system targets DNA regions associated with Plasmodium falciparum. The CRISPR-Cas12a enzyme recognizes the target sequence and uses fluorophore-quencher probes or FAM-Biotin probes in lateral flow assays, to generate signals with the help of specific guide RNAs (gRNAs). AI algorithms are used to design the gRNAs, ensuring low off-target effects and high specificity. The performance of the CRISPR system is further improved, and the prediction algorithms are continuously refined, by feeding diagnostic data into AI models. This integration combines the rapid and accurate detection capabilities of CRISPR with AI’s predictive power

Fig. 5

Strain distinction at single-base resolution is achievable with the NASBA-CRISPR cleavage (NASBACC) approach, as demonstrated by genotyping of the Zika virus. The activation of a toehold switch sensor is influenced by a synthetic trigger sequence, which is unique to strain variations after NASBA amplification of RNA. PAM sites specific to a strain generate either full-length or shortened RNA, affecting sensor activation. Sensor 32B can distinguish between dengue RNA and Zika strains but cannot differentiate between American and African strains of the Zika virus. To overcome this limitation, NASBACC exploits a single-base SNP that results in a PAM site exclusive to the American strain. This allows Cas9 to cleave the RNA of the American strain, generating shortened RNA that does not activate the sensor, while the RNA of the African strain remains intact and triggers the sensor. The NASBACC framework provides accurate genotypic data and has potential practical applications, as Cas9 is compatible with lyophilization

Fig. 6

AI algorithms enhance CRISPR-mediated point-of-care diagnosis. This figure illustrates the process of CRISPR-mediated point-of-care diagnostics. The first step involves nucleic acid extraction from patient samples. Next, disease biomarkers are targeted using CRISPR-Cas systems, such as Cas12 and Cas13a, which are guided by RNA sequences. The detection process generates signals through fluorescence or lateral flow readouts. Diagrammatic depiction of CRISPR-Cas-based diagnostic systems employing Cas12 and Cas13: a The figure shows CRISPR-Cas-based detection techniques using a lateral flow readout (lower panel) or a fluorescent probe (upper panel). Variants of CRISPR-Cas cleave single-stranded DNA (ssDNA) or RNA (ssRNA) that is bound to fluorophore-quencher or FAM-biotin reporters, after targeting specific pathogenic DNA or RNA sequences. FAM-tagged molecules are visible on paper strips, while fluorescence devices detect and quantify the emitted fluorescent signals. b This panel shows a methylene blue (MB) probe integrated into an electrochemical biosensor (E-CRISPR). The electric current decreases as a result of the enzyme cleaving the ssDNA linker of the MB electrochemical tag when CRISPR-Cas12 identifies the target sequence. c A magnet-assisted volumetric bar-chart chip (MAV-chip) combined with CRISPR-Cas12a and platinum nanoparticles (PtNP) is shown in this panel. When target DNA is present, Cas12 cleaves the neighboring ssDNA reporter that is connected to PtNP and a magnetic bead, in addition to the target sequence. PtNP facilitates the conversion of H₂O₂ to O₂. The quantity of target DNA is measured by the change in ink on the chip, caused by the liberation of O₂ gas