Designing sensitive viral diagnostics with machine learning

Generative and predictive neural networks for the design of functional RNA molecules

RNA is a remarkably versatile molecule that has been engineered for applications in therapeutics, diagnostics, and in vivo information-processing systems. However, the complex relationship between the sequence, structure, and function of RNA often necessitates extensive experimental screening of candidate sequences. Here we present a generalized, efficient neural network architecture that utilizes the sequence and structure of RNA molecules (SANDSTORM) to inform functional predictions across a diverse range of settings. We pair these predictive models with generative adversarial RNA design networks (GARDN), allowing the generative modelling of a diverse range of functional RNA molecules with targeted experimental attributes. This approach enables the design of novel sequence candidates that outperform those encountered during training or returned by classical thermodynamic algorithms, and can be deployed using as few as 384 example sequences. SANDSTORM and GARDN thus represent powerful new predictive and generative tools for the development of RNA molecules with improved function.

Accurate Prediction of CRISPR/Cas13a Guide Activity Using Feature Selection and Deep Learning

CRISPR/Cas13a serves as a key tool for nucleic acid tests; therefore, accurate prediction of its activity is essential for creating robust and sensitive diagnosis. In this study, we create a dual-branch neural network model that achieves high prediction accuracy and classification performance across two independent CRISPR/Cas13a data sets, outperforming previously published models relying solely on sequence features. The model integrates direct sequence encoding with descriptive features and yields 99 key descriptive features out of 1553, extracted through statistical analysis, which critically influence guide-target interactions and Cas13a guide activity. By employing Shapley Additive Explanations and Integrated Gradients for feature importance analysis, we show that sequence composition, mismatch type and frequency, and the protospacer flanking site region are primary features. These findings underscore the importance of using descriptive features as complementary inputs to deep learning-based encoding and provide valuable insights into the mechanisms underlying guide-target interaction. All in all, this study not only introduces a reliable and efficient model for Cas13a guide activity prediction but also offers a foundation for future rational design efforts.

A Streamlined Point-of-Care CRISPR Test for Tuberculosis Detection Directly from Sputum

Mycobacterium tuberculosis (Mtb) is a major threat to global health and is responsible for over one million deaths each year. To stem the tide of cases and maximize opportunities for early interventions, there is an urgent need for affordable and simple means of tuberculosis diagnosis in under-resourced areas. We sought to develop a CRISPR-based isothermal assay coupled with a compatible, straightforward sample processing technique for point-of-care use. Here, we combine Recombinase Polymerase Amplification (RPA) with Cas13a and Cas12a, to create two parallelised one-pot assays that detect two conserved elements of Mtb (IS6110 and IS1081) and an internal control targeting human DNA. These assays were shown to be compatible with lateral flow and can be readily lyophilized. Our finalized assay exhibited sensitivity over a wide range of bacterial loads (105 to 102 CFU/mL) in sputum. The limit of detection (LoD) of the assay was determined to be 69.0 (51.0 - 86.9) CFU/mL for Mtb strain H37Rv spiked in sputum and 80.5 (59.4 - 101.6) CFU/mL for M. bovis BCG. Our assay showed no cross reactivity against a wide range of bacterial/fungal isolates. Clinical tests on 13 blinded sputum samples revealed 100% (6/6) sensitivity and 100% (7/7) specificity compared to culture. Our assay exhibited comparable sensitivity in clinical samples to the microbiological gold standard, TB culture, and to the nucleic acid state-of-the-art, GeneXpert MTB/RIF Ultra. This technology streamlines TB diagnosis from sample extraction to assay readout in a rapid and robust format, making it the first test to combine amplification and detection while being compatible with both lateral flow and lyophilization.

Monitoring the Land and Sea: Enhancing Efficiency Through CRISPR-Cas Driven Depletion and Enrichment of Environmental DNA

Characterizing biodiversity using environmental DNA (eDNA) represents a paradigm shift in our capacity for biomonitoring complex environments, both aquatic and terrestrial. However, eDNA biomonitoring is limited by biases toward certain species and the low taxonomic resolution of current metabarcoding approaches. Shotgun metagenomics of eDNA enables the collection of whole ecosystem data by sequencing all molecules present, allowing characterization and identification. Clustered regularly interspaced short palindromic repeats (CRISPR) and the CRISPR-associated proteins (Cas)-based methods have the potential to improve the efficiency of eDNA metagenomic sequencing of low-abundant target organisms and simplify data analysis by enrichment of target species or nontarget DNA depletion before sequencing. Implementation of CRISPR-Cas in eDNA has been limited due to a lack of interest and support in the past. This perspective synthesizes current approaches of CRISPR-Cas to study underrepresented taxa and advocate for further application and optimization of depletion and enrichment methods of eDNA using CRISPR-Cas, holding promise for eDNA biomonitoring.

PathoGD: an integrative genomics approach to primer and guide RNA design for CRISPR-based diagnostics

Critical to the success of CRISPR-based diagnostic assays is the selection of a diagnostic target highly specific to the organism of interest, a process often requiring iterative cycles of manual selection, optimisation, and redesign. Here we present PathoGD, a bioinformatic pipeline for rapid and high-throughput design of RPA primers and gRNAs for CRISPR-Cas12a-based pathogen detection. PathoGD is fully automated, leverages publicly available sequences and is scalable to large datasets, allowing rapid continuous monitoring and validation of primer/gRNA sets to ensure ongoing assay relevance. We designed primers and gRNAs for five clinically relevant bacterial pathogens, and experimentally validated a subset of the designs for detecting Streptococcus pyogenes and/or Neisseria gonorrhoeae in assays with and without pre-amplification. We demonstrated high specificity of primers and gRNAs designed, with minimal off-target signal observed for all combinations. We anticipate PathoGD will be an important resource for assay design for current and emerging pathogens. PathoGD is available on GitHub at https://github.com/sjlow23/pathogd .

Prognostic prediction models for postoperative patients with stage I to III colorectal cancer based on machine learning

BACKGROUND

Colorectal cancer (CRC) is characterized by high heterogeneity, aggressiveness, and high morbidity and mortality rates. With machine learning (ML) algorithms, patient, tumor, and treatment features can be used to develop and validate models for predicting survival. In addition, important variables can be screened and different applications can be provided that could serve as vital references when making clinical decisions and potentially improving patient outcomes in clinical settings.

AIM

To construct prognostic prediction models and screen important variables for patients with stage I to III CRC.

METHODS

More than 1000 postoperative CRC patients were grouped according to survival time (with cutoff values of 3 years and 5 years) and assigned to training and testing cohorts (7:3). For each 3-category survival time, predictions were made by 4 ML algorithms (all-variable and important variable-only datasets), each of which was validated via 5-fold cross-validation and bootstrap validation. Important variables were screened with multivariable regression methods. Model performance was evaluated and compared before and after variable screening with the area under the curve (AUC). SHapley Additive exPlanations (SHAP) further demonstrated the impact of important variables on model decision-making. Nomograms were constructed for practical model application.

RESULTS

Our ML models performed well; the model performance before and after important parameter identification was consistent, and variable screening was effective. The highest pre- and postscreening model AUCs 95% confidence intervals in the testing set were 0.87 (0.81-0.92) and 0.89 (0.84-0.93) for overall survival, 0.75 (0.69-0.82) and 0.73 (0.64-0.81) for disease-free survival, 0.95 (0.88-1.00) and 0.88 (0.75-0.97) for recurrence-free survival, and 0.76 (0.47-0.95) and 0.80 (0.53-0.94) for distant metastasis-free survival. Repeated cross-validation and bootstrap validation were performed in both the training and testing datasets. The SHAP values of the important variables were consistent with the clinicopathological characteristics of patients with tumors. The nomograms were created.

CONCLUSION

We constructed a comprehensive, high-accuracy, important variable-based ML architecture for predicting the 3-category survival times. This architecture could serve as a vital reference for managing CRC patients.

Preamplification-free viral RNA diagnostics with single-nucleotide resolution using MARVE, an origami paper-based colorimetric nucleic acid test

The evolution and mutation of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) are urgent concerns as they pose the risk of vaccine failure and increased viral transmission. However, affordable and scalable tools allowing rapid identification of SARS-CoV-2 variants are not readily available, which impedes diagnosis and epidemiological surveillance. Here we present a colorimetric nucleic acid assay named MARVE (multiplexed, preamplification-free, single-nucleotide-resolved viral evolution) that is convenient to perform and yields single-nucleotide resolution. The assay integrates nucleic acid strand displacement reactions with enzymatic amplification to colorimetrically sense viral RNA using a metal ion-incorporated DNA probe (TEprobe). We provide detailed guidelines to design TEprobes for discriminating single-nucleotide variations in viral RNAs, and to fabricate a test paper for the detection of SARS-CoV-2 variants of concern. Compared with other nucleic acid assays, our assay is preamplification-free, single-nucleotide-resolvable and results are visible via a color change. Besides, it is smartphone readable, multiplexed, quick and cheap ($0.30 per test). The protocol takes ~2 h to complete, from the design and preparation of the DNA probes and test papers (~1 h) to the detection of SARS-CoV-2 or its variants (30-45 min). The design of the TEprobes requires basic knowledge of molecular biology and familiarity with NUPACK or the Python programming language. The fabrication of the origami papers requires access to a wax printer using the CAD and PDF files provided or requires users to be familiar with AutoCAD to design new origami papers. The protocol is also applicable for designing assays to detect other pathogens and their variants.

Model-directed generation of artificial CRISPR-Cas13a guide RNA sequences improves nucleic acid detection

CRISPR guide RNA sequences deriving exactly from natural sequences may not perform optimally in every application. Here we implement and evaluate algorithms for designing maximally fit, artificial CRISPR-Cas13a guides with multiple mismatches to natural sequences that are tailored for diagnostic applications. These guides offer more sensitive detection of diverse pathogens and discrimination of pathogen variants compared with guides derived directly from natural sequences and illuminate design principles that broaden Cas13a targeting.

CRISPR for companion diagnostics in low-resource settings

New point-of-care tests (POCTs), which are especially useful in low-resource settings, are needed to expand screening capacity for diseases that cause significant mortality: tuberculosis, multiple cancers, and emerging infectious diseases. Recently, clustered regularly interspaced short palindromic repeats (CRISPR)-based diagnostic (CRISPR-Dx) assays have emerged as powerful and versatile alternatives to traditional nucleic acid tests, revealing a strong potential to meet this need for new POCTs. In this review, we discuss CRISPR-Dx assay techniques that have been or could be applied to develop POCTs, including techniques for sample processing, target amplification, multiplex assay design, and signal readout. This review also describes current and potential applications for POCTs in disease diagnosis and includes future opportunities and challenges for such tests. These tests need to advance beyond initial assay development efforts to broadly meet criteria for use in low-resource settings.

Integrating allele-specific PCR with CRISPR-Cas13a for sensitive KRAS mutation detection in pancreatic cancer

BACKGROUND

The clustered regulatory interspaced short palindromic repeats (CRISPR)-Cas13a system has strong potential for highly sensitive detection of exogenous sequences. The detection of KRASG12 point mutations with low allele frequencies may prove powerful for the formal diagnosis of pancreatic ductal adenocarcinoma (PDAC).

RESULTS

We implemented preamplification of KRAS alleles (wild-type and mutant) to reveal the presence of mutant KRAS with CRISPR-Cas13a. The discrimination of KRASG12D from KRASWT was poor for the generic KRAS preamplification templates and depended on the crRNA design, the secondary structure of the target templates, and the nature of the mismatches between the guide and the templates. To improve the specificity, we used an allele-specific PCR preamplification method called CASPER (Cas13a Allele-Specific PCR Enzyme Recognition). CASPER enabled specific and sensitive detection of KRASG12D with low DNA input. CASPER detected KRAS mutations in DNA extracted from patients' pancreatic ultrasound-guided fine-needle aspiration fluid.

CONCLUSION

CASPER is easy to implement and is a versatile and reliable method that is virtually adaptable to any point mutation.

Precise pathogen quantification by CRISPR-Cas: a sweet but tough nut to crack

Pathogen detection is increasingly applied in medical diagnosis, food processing and safety, and environmental monitoring. Rapid, sensitive, and accurate pathogen quantification is the most critical prerequisite for assessing protocols and preventing risks. Among various methods evolved, those based on clustered regularly interspaced short palindromic repeats (CRISPR)-associated proteins (Cas) have been developed as important pathogen detection strategies due to their distinct advantages of rapid target recognition, programmability, ultra-specificity, and potential for scalability of point-of-care testing (POCT). However, arguments and concerns on the quantitative capability of CRISPR-based strategies are ongoing. Herein, we systematically overview CRISPR-based pathogen quantification strategies according to the principles, properties, and application scenarios. Notably, we review future challenges and perspectives to address the of precise pathogen quantification by CRISPR-Cas. We hope the insights presented in this review will benefit development of CRISPR-based pathogen detection methods.

CRISPR-Cas13: Pioneering RNA Editing for Nucleic Acid Therapeutics

The CRISPR-Cas13 system has emerged as a revolutionary tool for RNA editing, offering new opportunities for the development of nucleic acid therapeutics. Unlike DNA-targeting CRISPR-Cas9, Cas13 targets and cleaves RNA, enabling gene silencing and preventing genomic instability. Its applications include suppressing disease-causing genes, correcting splicing errors, and modulating immune responses. Despite these advances, challenges persist, such as the need to refine specificity, mitigate off-target impacts, and ensure effective delivery. This review provides an overview of the CRISPR-Cas13 mechanism, elucidating its role in RNA-targeted therapies and its transformative potential for disease treatment. Furthermore, it addresses the ongoing challenges that the scientific community is striving to overcome.

Therapeutic Application and Structural Features of Adeno-Associated Virus Vector

Adeno-associated virus (AAV) is characterized by non-pathogenicity, long-term infection, and broad tropism and is actively developed as a vector virus for gene therapy products. AAV is classified into more than 100 serotypes based on differences in the amino acid sequence of the capsid protein. Endocytosis involves the uptake of viral particles by AAV and accessory receptors during AAV infection. After entry into the cell, they are transported to the nucleus through the nuclear pore complex. AAVs mainly use proteoglycans as receptors to enter cells, but the types of sugar chains in proteoglycans that have binding ability are different. Therefore, it is necessary to properly evaluate the primary structure of receptor proteins, such as amino acid sequences and post-translational modifications, including glycosylation, and the higher-order structure of proteins, such as the folding of the entire capsid structure and the three-dimensional (3D) structure of functional domains, to ensure the efficacy and safety of biopharmaceuticals. To further enhance safety, it is necessary to further improve the efficiency of gene transfer into target cells, reduce the amount of vector administered, and prevent infection of non-target cells.

Deep learning enhancing guide RNA design for CRISPR/Cas12a-based diagnostics

Rapid and accurate diagnostic tests are fundamental for improving patient outcomes and combating infectious diseases. The Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) Cas12a-based detection system has emerged as a promising solution for on-site nucleic acid testing. Nonetheless, the effective design of CRISPR RNA (crRNA) for Cas12a-based detection remains challenging and time-consuming. In this study, we propose an enhanced crRNA design system with deep learning for Cas12a-mediated diagnostics, referred to as EasyDesign. This system employs an optimized convolutional neural network (CNN) prediction model, trained on a comprehensive data set comprising 11,496 experimentally validated Cas12a-based detection cases, encompassing a wide spectrum of prevalent pathogens, achieving Spearman's ρ = 0.812. We further assessed the model performance in crRNA design for four pathogens not included in the training data: Monkeypox Virus, Enterovirus 71, Coxsackievirus A16, and Listeria monocytogenes. The results demonstrated superior prediction performance compared to the traditional experiment screening. Furthermore, we have developed an interactive web server (https://crispr.zhejianglab.com/) that integrates EasyDesign with recombinase polymerase amplification (RPA) primer design, enhancing user accessibility. Through this web-based platform, we successfully designed optimal Cas12a crRNAs for six human papillomavirus (HPV) subtypes. Remarkably, all the top five predicted crRNAs for each HPV subtype exhibited robust fluorescent signals in CRISPR assays, thereby suggesting that the platform could effectively facilitate clinical sample testing. In conclusion, EasyDesign offers a rapid and reliable solution for crRNA design in Cas12a-based detection, which could serve as a valuable tool for clinical diagnostics and research applications.

A Review of Probe-Based Enrichment Methods to Inform Plant Virus Diagnostics

Modern diagnostic techniques based on DNA sequence similarity are currently the gold standard for the detection of existing and emerging pathogens. Whilst individual assays are inexpensive to use, assay development is costly and carries risks of not being sensitive or specific enough to capture an increasingly diverse range of targets. Sequencing can provide the entire nucleic acid content of a sample and may be used to identify all pathogens present in the sample when the depth of coverage is sufficient. Targeted enrichment techniques have been used to increase sequence coverage and improve the sensitivity of detection within virus samples, specifically, to capture sequences for a range of different viruses or increase the number of reads from low-titre virus infections. Vertebrate viruses have been well characterised using in-solution hybridisation capture to target diverse virus families. The use of probes for genotyping and strain identification has been limited in plants, and uncertainty around sensitivity is an impediment to the development of a large-scale virus panel to use within regulatory settings and diagnostic pipelines. This review aims to compare significant studies that have used targeted enrichment of viruses to identify approaches to probe design and potential for use in plant virus detection and characterisation.

CRISPR-Based Assays for Point-of-Need Detection and Subtyping of Influenza

The high disease burden of influenza virus poses a significant threat to human health. Optimized diagnostic technologies that combine speed, sensitivity, and specificity with minimal equipment requirements are urgently needed to detect the many circulating species, subtypes, and variants of influenza at the point of need. Here, we introduce such a method using Streamlined Highlighting of Infections to Navigate Epidemics (SHINE), a clustered regularly interspaced short palindromic repeats (CRISPR)-based RNA detection platform. Four SHINE assays were designed and validated for the detection and differentiation of clinically relevant influenza species (A and B) and subtypes (H1N1 and H3N2). When tested on clinical samples, these optimized assays achieved 100% concordance with quantitative RT-PCR. Duplex Cas12a/Cas13a SHINE assays were also developed to detect two targets simultaneously. This study demonstrates the utility of this duplex assay in discriminating two alleles of an oseltamivir resistance (H275Y) mutation as well as in simultaneously detecting influenza A and human RNAse P in patient samples. These assays have the potential to expand influenza detection outside of clinical laboratories for enhanced influenza diagnosis and surveillance.

A versatile automated pipeline for quantifying virus infectivity by label-free light microscopy and artificial intelligence

Virus infectivity is traditionally determined by endpoint titration in cell cultures, and requires complex processing steps and human annotation. Here we developed an artificial intelligence (AI)-powered automated framework for ready detection of virus-induced cytopathic effect (DVICE). DVICE uses the convolutional neural network EfficientNet-B0 and transmitted light microscopy images of infected cell cultures, including coronavirus, influenza virus, rhinovirus, herpes simplex virus, vaccinia virus, and adenovirus. DVICE robustly measures virus-induced cytopathic effects (CPE), as shown by class activation mapping. Leave-one-out cross-validation in different cell types demonstrates high accuracy for different viruses, including SARS-CoV-2 in human saliva. Strikingly, DVICE exhibits virus class specificity, as shown with adenovirus, herpesvirus, rhinovirus, vaccinia virus, and SARS-CoV-2. In sum, DVICE provides unbiased infectivity scores of infectious agents causing CPE, and can be adapted to laboratory diagnostics, drug screening, serum neutralization or clinical samples.

Electrochemical Sensors and Biosensors for Identification of Viruses: A Critical Review

Due to their life cycle, viruses can disrupt the metabolism of their hosts, causing diseases. If we want to disrupt their life cycle, it is necessary to identify their presence. For this purpose, it is possible to use several molecular-biological and bioanalytical methods. The reference selection was performed based on electronic databases (2020-2023). This review focused on electrochemical methods with high sensitivity and selectivity (53% voltammetry/amperometry, 33% impedance, and 12% other methods) which showed their great potential for detecting various viruses. Moreover, the aforementioned electrochemical methods have considerable potential to be applicable for care-point use as they are portable due to their miniaturizability and fast speed analysis (minutes to hours), and are relatively easy to interpret. A total of 2011 articles were found, of which 86 original papers were subsequently evaluated (the majority of which are focused on human pathogens, whereas articles dealing with plant pathogens are in the minority). Thirty-two species of viruses were included in the evaluation. It was found that most of the examined research studies (77%) used nanotechnological modifications. Other ones performed immunological (52%) or genetic analyses (43%) for virus detection. 5% of the reports used peptides to increase the method's sensitivity. When evaluable, 65% of the research studies had LOD values in the order of ng or nM. The vast majority (79%) of the studies represent proof of concept and possibilities with low application potential and a high need of further research experimental work.

Prediction of on-target and off-target activity of CRISPR-Cas13d guide RNAs using deep learning

Transcriptome engineering applications in living cells with RNA-targeting CRISPR effectors depend on accurate prediction of on-target activity and off-target avoidance. Here we design and test ~200,000 RfxCas13d guide RNAs targeting essential genes in human cells with systematically designed mismatches and insertions and deletions (indels). We find that mismatches and indels have a position- and context-dependent impact on Cas13d activity, and mismatches that result in G-U wobble pairings are better tolerated than other single-base mismatches. Using this large-scale dataset, we train a convolutional neural network that we term targeted inhibition of gene expression via gRNA design (TIGER) to predict efficacy from guide sequence and context. TIGER outperforms the existing models at predicting on-target and off-target activity on our dataset and published datasets. We show that TIGER scoring combined with specific mismatches yields the first general framework to modulate transcript expression, enabling the use of RNA-targeting CRISPRs to precisely control gene dosage.

New design strategies for ultra-specific CRISPR-Cas13a-based RNA detection with single-nucleotide mismatch sensitivity

An increasingly pressing need for clinical diagnostics has required the development of novel nucleic acid-based detection technologies that are sensitive, fast, and inexpensive, and that can be deployed at point-of-care. Recently, the RNA-guided ribonuclease CRISPR-Cas13 has been successfully harnessed for such purposes. However, developing assays for detection of genetic variability, for example single-nucleotide polymorphisms, is still challenging and previously described design strategies are not always generalizable. Here, we expanded our characterization of LbuCas13a RNA-detection specificity by performing a combination of experimental RNA mismatch tolerance profiling, molecular dynamics simulations, protein, and crRNA engineering. We found certain positions in the crRNA-target-RNA duplex that are particularly sensitive to mismatches and establish the effect of RNA concentration in mismatch tolerance. Additionally, we determined that shortening the crRNA spacer or modifying the direct repeat of the crRNA leads to stricter specificities. Furthermore, we harnessed our understanding of LbuCas13a allosteric activation pathways through molecular dynamics and structure-guided engineering to develop novel Cas13a variants that display increased sensitivities to single-nucleotide mismatches. We deployed these Cas13a variants and crRNA design strategies to achieve superior discrimination of SARS-CoV-2 strains compared to wild-type LbuCas13a. Together, our work provides new design criteria and Cas13a variants to use in future easier-to-implement Cas13-based RNA detection applications.

CRISPR-Cas system: A promising tool for rapid detection of SARS-CoV-2 variants

COVID-19, caused by SARS-CoV-2, remains a global health crisis. The emergence of multiple variants with enhanced characteristics necessitates their detection and monitoring. Genome sequencing, the gold standard, faces implementation challenges due to complexity, cost, and limited throughput. The CRISPR-Cas system offers promising potential for rapid variant detection, with advantages such as speed, sensitivity, specificity, and programmability. This review provides an in-depth examination of the applications of CRISPR-Cas in mutation detection specifically for SARS-CoV-2. It begins by introducing SARS-CoV-2 and existing variant detection platforms. The principles of the CRISPR-Cas system are then clarified, followed by an exploration of three CRISPR-Cas-based mutation detection platforms, which are evaluated from different perspectives. The review discusses strategies for mutation site selection and the utilization of CRISPR-Cas, offering valuable insights for the development of mutation detection methods. Furthermore, a critical analysis of the clinical applications, advantages, disadvantages, challenges, and prospects of the CRISPR-Cas system is provided.

Deep learning and CRISPR-Cas13d ortholog discovery for optimized RNA targeting

Effective and precise mammalian transcriptome engineering technologies are needed to accelerate biological discovery and RNA therapeutics. Despite the promise of programmable CRISPR-Cas13 ribonucleases, their utility has been hampered by an incomplete understanding of guide RNA design rules and cellular toxicity resulting from off-target or collateral RNA cleavage. Here, we quantified the performance of over 127,000 RfxCas13d (CasRx) guide RNAs and systematically evaluated seven machine learning models to build a guide efficiency prediction algorithm orthogonally validated across multiple human cell types. Deep learning model interpretation revealed preferred sequence motifs and secondary features for highly efficient guides. We next identified and screened 46 novel Cas13d orthologs, finding that DjCas13d achieves low cellular toxicity and high specificity-even when targeting abundant transcripts in sensitive cell types, including stem cells and neurons. Our Cas13d guide efficiency model was successfully generalized to DjCas13d, illustrating the power of combining machine learning with ortholog discovery to advance RNA targeting in human cells.

RNA structure modulates Cas13 activity and enables mismatch detection

The RNA-targeting CRISPR nuclease Cas13 has emerged as a powerful tool for applications ranging from nucleic acid detection to transcriptome engineering and RNA imaging1-6. Cas13 is activated by the hybridization of a CRISPR RNA (crRNA) to a complementary single-stranded RNA (ssRNA) protospacer in a target RNA1,7. Though Cas13 is not activated by double-stranded RNA (dsRNA) in vitro, it paradoxically demonstrates robust RNA targeting in environments where the vast majority of RNAs are highly structured2,8. Understanding Cas13's mechanism of binding and activation will be key to improving its ability to detect and perturb RNA; however, the mechanism by which Cas13 binds structured RNAs remains unknown9. Here, we systematically probe the mechanism of LwaCas13a activation in response to RNA structure perturbations using a massively multiplexed screen. We find that there are two distinct sequence-independent modes by which secondary structure affects Cas13 activity: structure in the protospacer region competes with the crRNA and can be disrupted via a strand-displacement mechanism, while structure in the region 3' to the protospacer has an allosteric inhibitory effect. We leverage the kinetic nature of the strand displacement process to improve Cas13-based RNA detection, enhancing mismatch discrimination by up to 50-fold and enabling sequence-agnostic mutation identification at low (<1%) allele frequencies. Our work sets a new standard for CRISPR-based nucleic acid detection and will enable intelligent and secondary-structure-guided target selection while also expanding the range of RNAs available for targeting with Cas13.

Development of Cas13a-based assays for Neisseria gonorrhoeae detection and gyrase A determination

Neisseria gonorrhoeae is one of the most common bacterial sexually transmitted infections. The emergence of antimicrobial-resistant N. gonorrhoeae is an urgent public health threat. Currently, the diagnosis of N. gonorrhoeae infection requires expensive laboratory infrastructure, while antimicrobial susceptibility determination requires bacterial culture, both of which are infeasible in low-resource areas where the prevalence of infection is highest. Recent advances in molecular diagnostics, such as specific high-sensitivity enzymatic reporter unlocking (SHERLOCK) using CRISPR-Cas13a and isothermal amplification, have the potential to provide low-cost detection of pathogen and antimicrobial resistance. We designed and optimized RNA guides and primer sets for SHERLOCK assays capable of detecting N. gonorrhoeae via the porA gene and of predicting ciprofloxacin susceptibility via a single mutation in the gyrase A (gyrA) gene. We evaluated their performance using both synthetic DNA and purified N. gonorrhoeae isolates. For porA, we created both a fluorescence-based assay and lateral flow assay using a biotinylated fluorescein reporter. Both methods demonstrated sensitive detection of 14 N. gonorrhoeae isolates and no cross-reactivity with 3 non-gonococcal Neisseria isolates. For gyrA, we created a fluorescence-based assay that correctly distinguished between 20 purified N. gonorrhoeae isolates with phenotypic ciprofloxacin resistance and 3 with phenotypic susceptibility. We confirmed the gyrA genotype predictions from the fluorescence-based assay with DNA sequencing, which showed 100% concordance for the isolates studied. We report the development of Cas13a-based SHERLOCK assays that detect N. gonorrhoeae and differentiate ciprofloxacin-resistant isolates from ciprofloxacin-susceptible isolates. IMPORTANCE Neisseria gonorrhoeae, the cause of gonorrhea, disproportionately affects resource-limited settings. Such areas, however, lack the technical capabilities for diagnosing the infection. The consequences of poor or absent diagnostics include increased disease morbidity, which, for gonorrhea, includes an increased risk for HIV infection, infertility, and neonatal blindness, as well as an overuse of antibiotics that contributes to the emergence of antibiotic resistance. We used a novel CRISPR-based technology to develop a rapid test that does not require laboratory infrastructure for both diagnosing gonorrhea and predicting whether ciprofloxacin can be used in its treatment, a one-time oral pill. With further development, that diagnostic test may be of use in low-resource settings.

Model-directed generation of CRISPR-Cas13a guide RNAs designs artificial sequences that improve nucleic acid detection

Generating maximally-fit biological sequences has the potential to transform CRISPR guide RNA design as it has other areas of biomedicine. Here, we introduce model-directed exploration algorithms (MEAs) for designing maximally-fit, artificial CRISPR-Cas13a guides-with multiple mismatches to any natural sequence-that are tailored for desired properties around nucleic acid diagnostics. We find that MEA-designed guides offer more sensitive detection of diverse pathogens and discrimination of pathogen variants compared to guides derived directly from natural sequences, and illuminate interpretable design principles that broaden Cas13a targeting.

Highly multiplexed mRNA quantitation with CRISPR-Cas13

RNA quantitation tools are often either high-throughput or cost-effective, but rarely are they both. Existing methods can profile the transcriptome at great expense or are limited to quantifying a handful of genes by labor constraints. A technique that permits more throughput at a reduced cost could enable multi-gene kinetic studies, gene regulatory network analysis, and combinatorial genetic screens. Here, we introduce quantitative Combinatorial Arrayed Reactions for Multiplexed Evaluation of Nucleic acids (qCARMEN): an RNA quantitation technique which leverages the programmable RNA-targeting capabilities of CRISPR-Cas13 to address this challenge by quantifying over 4,500 gene-sample pairs in a single experiment. Using qCARMEN, we studied the response profiles of interferon-stimulated genes (ISGs) during interferon (IFN) stimulation and flavivirus infection. Additionally, we observed isoform switching kinetics during epithelial-mesenchymal transition. qCARMEN is a simple and inexpensive technique that greatly enhances the scalability of RNA quantitation for novel applications with performance similar to gold-standard methods.

CRISPR-Cas-Based Biomonitoring for Marine Environments: Toward CRISPR RNA Design Optimization Via Deep Learning

Almost all of Earth's oceans are now impacted by multiple anthropogenic stressors, including the spread of nonindigenous species, harmful algal blooms, and pathogens. Early detection is critical to manage these stressors effectively and to protect marine systems and the ecosystem services they provide. Molecular tools have emerged as a promising solution for marine biomonitoring. One of the latest advancements involves utilizing CRISPR-Cas technology to build programmable, rapid, ultrasensitive, and specific diagnostics. CRISPR-based diagnostics (CRISPR-Dx) has the potential to allow robust, reliable, and cost-effective biomonitoring in near real time. However, several challenges must be overcome before CRISPR-Dx can be established as a mainstream tool for marine biomonitoring. A critical unmet challenge is the need to design, optimize, and experimentally validate CRISPR-Dx assays. Artificial intelligence has recently been presented as a potential approach to tackle this challenge. This perspective synthesizes recent advances in CRISPR-Dx and machine learning modeling approaches, showcasing CRISPR-Dx potential to progress as a rising molecular tool candidate for marine biomonitoring applications.

New design strategies for ultra-specific CRISPR-Cas13a-based RNA-diagnostic tools with single-nucleotide mismatch sensitivity

The pressing need for clinical diagnostics has required the development of novel nucleic acid-based detection technologies that are sensitive, fast, and inexpensive, and that can be deployed at point-of-care. Recently, the RNA-guided ribonuclease CRISPR-Cas13 has been successfully harnessed for such purposes. However, developing assays for detection of genetic variability, for example single-nucleotide polymorphisms, is still challenging and previously described design strategies are not always generalizable. Here, we expanded our characterization of LbuCas13a RNA-detection specificity by performing a combination of experimental RNA mismatch tolerance profiling, molecular dynamics simulations, protein, and crRNA engineering. We found certain positions in the crRNA-target-RNA duplex that are particularly sensitive to mismatches and establish the effect of RNA concentration in mismatch tolerance. Additionally, we determined that shortening the crRNA spacer or modifying the direct repeat of the crRNA leads to stricter specificities. Furthermore, we harnessed our understanding of LbuCas13a allosteric activation pathways through molecular dynamics and structure-guided engineering to develop novel Cas13a variants that display increased sensitivities to single-nucleotide mismatches. We deployed these Cas13a variants and crRNA design strategies to achieve superior discrimination of SARS-CoV-2 strains compared to wild-type LbuCas13a. Together, our work provides new design criteria and new Cas13a variants for easier-to-implement Cas13-based diagnostics.

KEY POINTS

Certain positions in the Cas13a crRNA-target-RNA duplex are particularly sensitive to mismatches.Understanding Cas13a's allosteric activation pathway allowed us to develop novel high-fidelity Cas13a variants.These Cas13a variants and crRNA design strategies achieve superior discrimination of SARS-CoV-2 strains.

GRAPHICAL ABSTRACT

Leveraging artificial intelligence in the fight against infectious diseases

Despite advances in molecular biology, genetics, computation, and medicinal chemistry, infectious disease remains an ominous threat to public health. Addressing the challenges posed by pathogen outbreaks, pandemics, and antimicrobial resistance will require concerted interdisciplinary efforts. In conjunction with systems and synthetic biology, artificial intelligence (AI) is now leading to rapid progress, expanding anti-infective drug discovery, enhancing our understanding of infection biology, and accelerating the development of diagnostics. In this Review, we discuss approaches for detecting, treating, and understanding infectious diseases, underscoring the progress supported by AI in each case. We suggest future applications of AI and how it might be harnessed to help control infectious disease outbreaks and pandemics.

Generative and predictive neural networks for the design of functional RNA molecules

RNA is a remarkably versatile molecule that has been engineered for applications in therapeutics, diagnostics, and in vivo information-processing systems. However, the complex relationship between the sequence and structural properties of an RNA molecule and its ability to perform specific functions often necessitates extensive experimental screening of candidate sequences. Here we present a generalized neural network architecture that utilizes the sequence and structure of RNA molecules (SANDSTORM) to inform functional predictions. We demonstrate that this approach achieves state-of-the-art performance across several distinct RNA prediction tasks, while learning interpretable abstractions of RNA secondary structure. We paired these predictive models with generative adversarial RNA design networks (GARDN), allowing the generative modelling of novel mRNA 5' untranslated regions and toehold switch riboregulators exhibiting a predetermined fitness. This approach enabled the design of novel toehold switches with a 43-fold increase in experimentally characterized dynamic range compared to those designed using classic thermodynamic algorithms. SANDSTORM and GARDN thus represent powerful new predictive and generative tools for the development of diagnostic and therapeutic RNA molecules with improved function.

Development of Cas13a-based Assays for Neisseria gonorrhoeae Detection and Gyrase A Determination

Background

Neisseria gonorrhoeae is one of the most common bacterial sexually transmitted infections. The emergence of antimicrobial-resistant N. gonorrhoeae is an urgent public health threat. Currently, diagnosis of N. gonorrhoeae infection requires expensive laboratory infrastructure, while antimicrobial susceptibility determination requires bacterial culture, both of which are infeasible in low-resource areas where prevalence is highest. Recent advances in molecular diagnostics, such as Specific High-sensitivity Enzymatic Reporter unLOCKing (SHERLOCK) using CRISPR-Cas13a and isothermal amplification, have the potential to provide low-cost detection of pathogen and antimicrobial resistance.

Methods and Results

We designed and optimized RNA guides and primer-sets for SHERLOCK assays capable of detecting N. gonorrhoeae via the por A gene and of predicting ciprofloxacin susceptibility via a single mutation in the gyrase A ( gyr A) gene. We evaluated their performance using both synthetic DNA and purified N. gonorrhoeae isolates. For por A, we created both a fluorescence-based assay and lateral flow assay using a biotinylated FAM reporter. Both methods demonstrated sensitive detection of 14 N. gonorrhoeae isolates and no cross-reactivity with 3 non-gonococcal Neisseria isolates. For gyr A, we created a fluorescence-based assay that correctly distinguished between 20 purified N. gonorrhoeae isolates with phenotypic ciprofloxacin resistance and 3 with phenotypic susceptibility. We confirmed the gyr A genotype predictions from the fluorescence-based assay with DNA sequencing, which showed 100% concordance for the isolates studied.

Conclusion

We report the development of Cas13a-based SHERLOCK assays that detect N. gonorrhoeae and differentiate ciprofloxacin-resistant isolates from ciprofloxacin-susceptible isolates.

Machine learning for design of degenerate Cas13a crRNAs using lassa virus as a model of highly variable RNA target

The design of minimum CRISPR RNA (crRNA) sets for detection of diverse RNA targets using sequence degeneracy has not been systematically addressed. We tested candidate degenerate Cas13a crRNA sets designed for detection of diverse RNA targets (Lassa virus). A decision tree machine learning (ML) algorithm (RuleFit) was applied to define the top attributes that determine the specificity of degenerate crRNAs to elicit collateral nuclease activity. Although the total number of mismatches (0-4) is important, the specificity depends as well on the spacing of mismatches, and their proximity to the 5' end of the spacer. We developed a predictive algorithm for design of candidate degenerate crRNA sets, allowing improved discrimination between "included" and "excluded" groups of related target sequences. A single degenerate crRNA set adhering to these rules detected representatives of all Lassa lineages. Our general ML approach may be applied to the design of degenerate crRNA sets for any CRISPR/Cas system.

CRISPR techniques and potential for the detection and discrimination of SARS-CoV-2 variants of concern

The continuing evolution of the SARS-CoV-2 virus has led to the emergence of many variants, including variants of concern (VOCs). CRISPR-Cas systems have been used to develop techniques for the detection of variants. These techniques have focused on the detection of variant-specific mutations in the spike protein gene of SARS-CoV-2. These sequences mostly carry single-nucleotide mutations and are difficult to differentiate using a single CRISPR-based assay. Here we discuss the specificity of the Cas9, Cas12, and Cas13 systems, important considerations of mutation sites, design of guide RNA, and recent progress in CRISPR-based assays for SARS-CoV-2 variants. Strategies for discriminating single-nucleotide mutations include optimizing the position of mismatches, modifying nucleotides in the guide RNA, and using two guide RNAs to recognize the specific mutation sequence and a conservative sequence. Further research is needed to confront challenges in the detection and differentiation of variants and sublineages of SARS-CoV-2 in clinical diagnostic and point-of-care applications.

Capturing nucleic acid variants with precision using CRISPR diagnostics

CRISPR/Cas systems have the ability to precisely target nucleotide sequences and enable their rapid identification and modification. While nucleotide modification has enabled the therapeutic correction of diseases, the process of identifying the target DNA or RNA has greatly expanded the field of molecular diagnostics in recent times. CRISPR-based DNA/RNA detection through programmable nucleic acid binding or cleavage has been demonstrated for a large number of pathogenic and non-pathogenic targets. Combining CRISPR detection with nucleic acid amplification and a terminal signal readout step allowed the development of numerous rapid and robust nucleic acid platforms. Wherever the Cas effector can faithfully distinguish nucleobase variants in the target, the platform can also be extended for sequencing-free rapid variant detection. Some initial PAM disruption-based SNV detection reports were limited to finding or integrating mutated/mismatched nucleotides within the PAM sequences. In this review, we try to summarize the developments made in CRISPR diagnostics (CRISPRDx) to date emphasizing CRISPR-based SNV detection. We also discuss the applications where such diagnostic modalities can be put to use, covering various fields of clinical research, SNV screens, disease genotyping, primary surveillance during microbial infections, agriculture, food safety, and industrial biotechnology. The ease of rapid design and implementation of such multiplexable assays can potentially expand the applications of CRISPRDx in the domain of affinity-based target sequencing, with immense possibilities for low-cost, quick, and widespread usage. In the end, in combination with proximity assays and a suicidal gene approach, CRISPR-based in vivo SNV detection and cancer cell targeting can be formulated as personalized gene therapy.

Simplified Cas13-based assays for the fast identification of SARS-CoV-2 and its variants

The widespread transmission and evolution of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) call for rapid nucleic acid diagnostics that are easy to use outside of centralized clinical laboratories. Here we report the development and performance benchmarking of Cas13-based nucleic acid assays leveraging lyophilised reagents and fast sample inactivation at ambient temperature. The assays, which we named SHINEv.2 (for 'streamlined highlighting of infections to navigate epidemics, version 2'), simplify the previously reported RNA-extraction-free SHINEv.1 technology by eliminating heating steps and the need for cold storage of the reagents. SHINEv.2 detected SARS-CoV-2 in nasopharyngeal samples with 90.5% sensitivity and 100% specificity (benchmarked against the reverse transcription quantitative polymerase chain reaction) in less than 90 min, using lateral-flow technology and incubation in a heat block at 37 °C. SHINEv.2 also allows for the visual discrimination of the Alpha, Beta, Gamma, Delta and Omicron SARS-CoV-2 variants, and can be run without performance losses by using body heat. Accurate, easy-to-use and equipment-free nucleic acid assays could facilitate wider testing for SARS-CoV-2 and other pathogens in point-of-care and at-home settings.

Recent Advances in Clustered Regularly Interspaced Short Palindromic Repeats-Based Biosensors for Point-of-Care Pathogen Detection

Infectious pathogens are pressing concerns due to their heavy toll on global health and socioeconomic infrastructure. Rapid, sensitive, and specific pathogen detection methods are needed more than ever to control disease spreading. The fast evolution of clustered regularly interspaced short palindromic repeats (CRISPR)-based diagnostics (CRISPR-Dx) has opened a new horizon in the field of molecular diagnostics. This review highlights recent efforts in configuring CRISPR technology as an efficient diagnostic tool for pathogen detection. It starts with a brief introduction of different CRISPR-Cas effectors and their working principles for disease diagnosis. It then focuses on the evolution of laboratory-based CRISPR technology toward a potential point-of-care test, including the development of new signaling mechanisms, elimination of preamplification and sample pretreatment steps, and miniaturization of CRISPR reactions on digital assay chips and lateral flow devices. In addition, promising examples of CRISPR-Dx for pathogen detection in various real samples, such as blood, saliva, nasal swab, plant, and food samples, are highlighted. Finally, the challenges and perspectives of future development of CRISPR-Dx for infectious disease monitoring are discussed.

Rapid, adaptable and sensitive Cas13-based COVID-19 diagnostics using ADESSO

During the ongoing COVID-19 pandemic, PCR testing and antigen tests have proven critical for helping to stem the spread of its causative agent, SARS-CoV-2. However, these methods suffer from either general applicability and/or sensitivity. Moreover, the emergence of variant strains creates the need for flexibility to correctly and efficiently diagnose the presence of substrains. To address these needs we developed the diagnostic test ADESSO (Accurate Detection of Evolving SARS-CoV-2 through SHERLOCK (Specific High Sensitivity Enzymatic Reporter UnLOCKing) Optimization) which employs Cas13 to diagnose patients in 1 h without sophisticated equipment. Using an extensive panel of clinical samples, we demonstrate that ADESSO correctly identifies infected individuals at a sensitivity and specificity comparable to RT-qPCR on extracted RNA and higher than antigen tests for unextracted samples. Altogether, ADESSO is a fast, sensitive and cheap method that can be applied in a point of care setting to diagnose COVID-19 and can be quickly adjusted to detect new variants.

Multiplexed CRISPR-based microfluidic platform for clinical testing of respiratory viruses and identification of SARS-CoV-2 variants

The coronavirus disease 2019 (COVID-19) pandemic has demonstrated a clear need for high-throughput, multiplexed and sensitive assays for detecting severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and other respiratory viruses and their emerging variants. Here, we present a cost-effective virus and variant detection platform, called microfluidic Combinatorial Arrayed Reactions for Multiplexed Evaluation of Nucleic acids (mCARMEN), which combines CRISPR-based diagnostics and microfluidics with a streamlined workflow for clinical use. We developed the mCARMEN respiratory virus panel to test for up to 21 viruses, including SARS-CoV-2, other coronaviruses and both influenza strains, and demonstrated its diagnostic-grade performance on 525 patient specimens in an academic setting and 166 specimens in a clinical setting. We further developed an mCARMEN panel to enable the identification of 6 SARS-CoV-2 variant lineages, including Delta and Omicron, and evaluated it on 2,088 patient specimens with near-perfect concordance to sequencing-based variant classification. Lastly, we implemented a combined Cas13 and Cas12 approach that enables quantitative measurement of SARS-CoV-2 and influenza A viral copies in samples. The mCARMEN platform enables high-throughput surveillance of multiple viruses and variants simultaneously, enabling rapid detection of SARS-CoV-2 variants.

Current progress and open challenges for applying deep learning across the biosciences

Deep Learning (DL) has recently enabled unprecedented advances in one of the grand challenges in computational biology: the half-century-old problem of protein structure prediction. In this paper we discuss recent advances, limitations, and future perspectives of DL on five broad areas: protein structure prediction, protein function prediction, genome engineering, systems biology and data integration, and phylogenetic inference. We discuss each application area and cover the main bottlenecks of DL approaches, such as training data, problem scope, and the ability to leverage existing DL architectures in new contexts. To conclude, we provide a summary of the subject-specific and general challenges for DL across the biosciences.

Multiplexed detection of bacterial nucleic acids using Cas13 in droplet microarrays

Rapid and accurate diagnosis of infections is fundamental to individual patient care and public health management. Nucleic acid detection methods are critical to this effort, but are limited either in the breadth of pathogens targeted or by the expertise and infrastructure required. We present here a high-throughput system that enables rapid identification of bacterial pathogens, bCARMEN, which utilizes: (1) modular CRISPR-Cas13-based nucleic acid detection with enhanced sensitivity and specificity; and (2) a droplet microfluidic system that enables thousands of simultaneous, spatially multiplexed detection reactions at nanoliter volumes; and (3) a novel preamplification strategy that further enhances sensitivity and specificity. We demonstrate bCARMEN is capable of detecting and discriminating 52 clinically relevant bacterial species and several key antibiotic resistance genes. We further develop a simple proof of principle workflow using stabilized reagents and cell phone camera optical readout, opening up the possibility of a rapid point-of-care multiplexed bacterial pathogen identification and antibiotic susceptibility testing.