Computational Optimization of CRISPR-Cas13a sgRNAs Targeting the SARS-CoV-2 Spike Gene for SHERLOCK-Based Diagnostics

Abstract

Background:

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has triggered a global health crisis, emphasizing the urgent need for accurate and rapid diagnostic tools. Modern molecular biology technologies, including CRISPR-Cas systems, provide highly efficient strategies for viral detection. Bioinformatic pipelines are essential for identifying conserved genomic regions and enabling rational single-guide RNA (sgRNA) design.

Methods:

This study aimed to design specific sgRNAs targeting the spike gene of SARS-CoV-2 isolates from Iranian patients using the SHERLOCK diagnostic platform. Complete genomes of the RefSeq virus and 470 SARS-CoV-2 isolates, representing all variants of concern (VOCs) detected in Iran, were retrieved from the NCBI and GISAID databases. Multiple sequence alignment with ClustalW identified conserved sequences within the receptor-binding domain (RBD) that differ from the RBD of SARS-CoV and MERS-CoV RefSeq genomes. Based on these regions, sgRNAs and isothermal amplification primers were designed using ADAPT, OLIGO7, and the UCSC Genome Browser to maximize diagnostic sensitivity and specificity. Secondary and tertiary structures of sgRNA-target complexes were analyzed via RNAfold and RNAup to select the most efficient sgRNA–amplicon combination.

Results:

Twenty-two–nucleotide sgRNA candidates were initially selected based on sequence alignment, showing high similarity to the SARS-CoV-2 RefSeq and low homology to SARS-CoV and MERS-CoV genomes. Analyses of secondary structures, RNA–RNA interactions, and free energy identified 6 sgRNAs with favorable 2-dimensional conformations and strong interaction profiles. Among these, the sgRNA1–Amplicon2 sequence exhibited the most stable 3-dimensional structure and a molecular docking score of −309.67, indicating high sensitivity and specificity for viral detection.

Conclusion:

This study successfully designed an sgRNA with high sensitivity and specificity for rapid SARS-CoV-2 detection using the CRISPR-Cas13a system, informed by genomic analysis of Iranian isolates. The proposed approach provides an efficient framework for the rapid design and deployment of CRISPR-based diagnostic tools applicable to diverse viral pathogens.

Introduction

The global spread of SARS-CoV-2 has posed a substantial threat to public health and socio-economic stability. However, existing diagnostic and therapeutic strategies remain insufficient to fully contain and manage the pandemic. For instance, RT-PCR, considered the gold standard for SARS-CoV-2 detection, has limited accuracy in reliably identifying newly emerging SARS-CoV-2 variants, such as the Omicron VoC. ¹ In this context, CRISPR-Cas systems have emerged as innovative and complementary tools to conventional approaches. Widely adopted in biomedical research, CRISPR-based technologies are being applied in pathogen detection, antiviral therapy development, drug screening, and vaccine research. ² These systems offer promising solutions for the diagnosis, prevention, and treatment of SARS-CoV-2 and other emerging infectious diseases.

SARS-CoV-2, a positive-sense RNA virus in the Betacoronavirus genus, shares high genetic similarity with SARS-CoV and uses the angiotensin-converting enzyme 2 (ACE2) receptor for host cell entry. ³ Its ~30 kb genome encodes several structural and non-structural proteins, among which the spike (S) glycoprotein plays a pivotal role in receptor binding and membrane fusion. ⁴ The receptor-binding domain (RBD) of the S1 subunit undergoes dynamic conformational changes, shifting between “down” (receptor-inaccessible) and “up” (receptor-accessible) states, facilitating interaction with the ACE2 receptor. ⁵ Understanding the structural and functional properties of the spike protein is thus essential for developing neutralizing antibodies, vaccines, and precise diagnostic assays. ⁶

CRISPR-based diagnostics, particularly platforms utilizing Cas13a nucleases, have shown great potential in achieving rapid and specific detection of RNA viruses. Recent studies have demonstrated the utility of CRISPR-Cas systems for identifying a broad spectrum of viral pathogens in both cell-based and in vivo models. ⁷ Their integration into field-deployable diagnostic workflows has accelerated pathogen surveillance, clinical triage, and containment strategies.

Compared to traditional RT-PCR methods, CRISPR-based assays offer simplified workflows and reduced time-to-result. Nevertheless, when the viral load is low (eg, <10 nM), the sensitivity of CRISPR detection decreases significantly. ⁸ To address this limitation, most CRISPR workflows incorporate a nucleic acid amplification step—such as polymerase chain reaction (PCR), loop-mediated isothermal amplification (LAMP), or recombinase polymerase amplification (RPA)—prior to Cas-mediated detection. ⁹ A typical Cas13-based diagnostic pipeline thus includes sample collection, RNA extraction, amplification (eg, RT-RPA or RT-LAMP), and detection using Cas13a, with visual output via fluorescence or colorimetric readout. ¹⁰

Given the limitations of existing methods and the high potential of CRISPR-based systems, this study aimed to develop a novel Cas13a-based diagnostic platform for the rapid and precise identification of SARS-CoV-2 VOCs which utilizes RPA for amplification and is centered on a novel, optimized sgRNA design. We hypothesized that computationally driven, rational design of single-guide RNA (sgRNA) would yield molecules with enhanced binding affinity and specificity for target variant sequences. Consequently, this work focuses on the in-silico design and comprehensive bioinformatic evaluation of highly specific sgRNAs, targeting conserved and variant-specific regions within the Spike gene of key VOCs.

Methods

Rapid Detection and Sequence Preparation

Accurate, rapid, and cost-effective detection of nucleic acids facilitates the precise identification of pathogens, genotyping, and disease monitoring. In this applied study, all available SARS-CoV-2 genome sequences infecting the human population up to the date of analysis were retrieved from the NCBI database (https://www.ncbi.nlm.nih.gov/). These sequences were aligned using BioEdit software (https://bioedit.software.informer.com/), and a conserved region within the receptor-binding domain (RBD) of the viral genome was selected for designing both single guide RNAs (sgRNAs) and recombinase polymerase amplification (RPA) primers.

Using Oligo-7 primer design software, primers were developed to amplify and detect the selected target region. Among various evaluated amplification methods, RPA was selected due to its high sensitivity. The amplified DNA was transcribed into RNA via T7 RNA polymerase to allow detection with LwCas13a. This workflow—comprising RPA amplification, transcription by T7 polymerase, Cas13a-mediated RNA detection, and signal generation through RNA cleavage—is collectively referred to as the SHERLOCK method.

Data Collection and Sequence Acquisition

In this study, spike gene RBD sequences from 3 coronavirus species were retrieved from the NCBI database. The RBD regions of SARS-CoV and MERS-CoV were also included for comparative analysis. Genomic sequences of various human-infecting SARS-CoV-2 isolates were obtained from the GISAID database (www.gisaid.org), which provides open access to influenza and coronavirus genomic data, including those related to the COVID-19 pandemic.

Sequence Alignment and Analysis

To identify conserved regions and locate mutation hotspots within the RBD, multiple sequence alignments were performed between Iranian SARS-CoV-2 isolates and the reference genome using the CLUSTALW algorithm integrated into BioEdit software. ¹¹ Pairwise alignments were conducted to identify unique sequence motifs in the SARS-CoV-2 S-RBD, distinguishing it from the SARS-CoV and MERS-CoV spike RBDs. The alignments were analyzed in BioEdit to detect conserved domains and single nucleotide variations (SNVs).

Identification of Target Regions for sgRNA and RPA Primer Design

Conserved regions suitable for sgRNA and RPA primer design were identified using BioEdit. Additionally, a neighbor-joining (NJ) phylogenetic analysis of SARS-CoV-2 isolates from Iranian patients was performed using MEGA X software. ¹² Genetic diversity and mutational entropy were visualized using Nextclade–Nextstrain (https://clades.nextstrain.org/), which aligned the reference genome against the Iranian isolates to generate phylogenetic trees and entropy plots.

Among the aligned sequences, several highly conserved segments within the RBD were identified based on entropy analysis. Entropy values were calculated using BioEdit and MEGA X, with lower values indicating greater sequence conservation.

Design of sgRNA and Computational Analysis

A critical consideration in sgRNA design for Cas13-based diagnostics is the avoidance of overlap with RPA primers, which can lead to nonspecific amplification and increased background signals. sgRNAs were designed to specifically recognize SARS-CoV-2 genomic regions, or alternatively, to broadly detect related coronaviruses.

Initial sgRNA design was carried out using the ADAPT web server. ¹³ Candidate regions were also retrieved from Synthego’s ICE platform (https://ice.editco.bio/) and the UCSC Genome Browser (https://www.ucsc.edu/). These were compared with ADAPT-generated sequences to identify overlapping regions and optimize target selection.

Candidate sgRNAs were filtered to identify unique 22-nucleotide sequences across the SARS-CoV-2 genome. Their specificity was confirmed by ensuring the absence of complementary binding within the human transcriptome.

BLAST-Based Off-Target Assessment

To ensure target specificity, the selected sgRNA sequences were analyzed using Nucleotide BLAST (https://blast.ncbi.nlm.nih.gov/). The sequences were screened against the SARS-CoV and MERS-CoV genomes to confirm the absence of homology with related viruses. Additional BLAST analyses against the human genome were conducted to eliminate potential off-target effects.

RPA Primer Design

The positions and lengths of RPA primers were defined based on the selected sgRNA target sites. Forward and reverse primers were designed to flank the upstream and downstream regions of the sgRNA target to ensure complete amplification. OLIGO software (https://www.oligo.net/) was used to assess primer characteristics, including melting temperature (Tm), potential for hairpin formation, primer length, and predicted amplicon size.

RNA-RNA Interaction Prediction

The ViennaRNA web server (http://rna.tbi.univie.ac.at/) was used for RNA secondary structure analysis and prediction of RNA-RNA interactions. The RNAfold module calculated minimum free energy (MFE) structures using dynamic programing algorithms. Additionally, the partition function (PF) approach based on McCaskill’s algorithm was used to compute base-pairing probabilities. ¹⁴

To model binding accessibility and energetics between sgRNAs and their targets, RNAup was used. It estimates duplex interaction energies and binding region accessibility, offering a biophysically realistic assessment. ¹⁵

3D Structure Prediction and RNA-RNA Docking Using HNADOCK

The HNADOCK server was employed for 3D modeling and docking of RNA-RNA complexes. It utilizes an FFT-based sampling algorithm to predict binding conformations, which are scored using an internal function. 3D RNA structures were modeled based on input sequences and secondary structure constraints. Docking scores and RNA-RNA interface metrics were used to identify the most stable complexes.^16,17

Results

Data Collection and Sequence Preparation

The reference receptor-binding domain (RBD) sequences of SARS-CoV-2, SARS-CoV, and MERS-CoV genomes were retrieved from the NCBI database. In the reference genome of SARS-CoV-2, the RBD region spans nucleotides 22 516 to 23 184. Complete genome sequences of SARS-CoV-2 isolates from Iranian patients were downloaded from the GISAID database (https://gisaid.org/) and saved in FASTA format. As of May 10, 2022, approximately 226 735 SARS-CoV-2 sequences were available globally. By applying the filters “Asia” (continent), “Iran” (geographic region), “Complete genome,” “High coverage,” and “Patient status,” the dataset was refined to 470 high-quality sequences, which were used for subsequent analyses. Sequences with missing nucleotides, low coverage, or incomplete metadata were excluded to ensure data quality and reliability for downstream sgRNA design and bioinformatic analyses.

Identification of Conserved RBD Regions for sgRNA Design

As shown in Table 1, several highly conserved segments were identified and prioritized for sgRNA and RPA primer design. These regions exhibited exceptionally low entropy values, reflecting minimal sequence variability across the Iranian SARS-CoV-2 genomes. In Table 1, the Average entropy (Hx) column quantifies sequence diversity, with values near zero indicating high conservation. The Segment length refers to the nucleotide span of each conserved region, while the Consensus sequence denotes the predominant nucleotide composition. The Genomic position column indicates the coordinates based on the SARS-CoV-2 reference genome (Wuhan-Hu-1, NC_045512).

Table 1.

Conserved Segments With the Lowest Single Nucleotide Variation (SNV) Frequencies, Including Genomic Positions, Segment Lengths, Average Entropy (Hx), and Consensus Sequences.

Position	Consensus	Segment length	Average entropy (Hx)
22 526-22 586	AGAGTCCAACCAACAGAATCTATTGTTAGATTTCCTAATATTACAAACTTGTGCCCTTTTG	61	0.0028
22 715-22 783	GTGTCTCCTACTAAATTAAATGATCTCTGCTTTACTAATGTCTATGCAGATTCATTTGTAATTAGAGGT	69	0.0006
22 796-22 890	CAAATCGCTCCAGGGCAAA	26	0.0105
22 823-22 890	ATTGCTGATTATAATTATAAATTACCAGATGATTTTACAGGCTGCGTTATAGCTTGGAATTCTAACAA	68	0.0000
22 892-22 906	CTTGATTCTAAGGTT	15	0.0000
22 908-22 925	GTGGTAATTATAATTACC	18	0.0030
22 927-22 961	GTATAGATTGTTTAGGAAGTCTAATCTCAAACCTT	35	0.0013
23 023-23 048	AGGTTTTAATTGTTACTTTCCTTTAC	26	0.0035
23 092-23 210	ATACAGAGTAGTAGTACTTTCTTTTGAACTTCTACATGCACCAGCAACTGTTTGTGGACCTAAAAAGTCTACTAATTTGGTTAAAAACAAATGTGTCAATTTCAACTTCAATGGTTTAA	119	0.0018

Notably, the region spanning positions 22 823 to 22 890 exhibited the lowest entropy value of 0, indicating complete conservation among all analyzed isolates. These loci were subsequently selected as the basis for sgRNA and RPA primer development for CRISPR-based detection.

To comprehensively assess sequence conservation and genetic diversity, a composite figure was constructed (Figure 1), integrating 3 layers of analysis: (1) a Shannon entropy plot illustrating positional variability across aligned SARS-CoV-2 genomes, (2) a mutation frequency graph depicting the distribution of variants across the viral genome, and (3) a phylogenetic tree representing the relationships among all 470 Iranian isolates.

Figure 1.

Composite figure summarizing sequence conservation, phylogenetic diversity, and mutation frequency among SARS-CoV-2 isolates from Iranian patients. (A) Shannon entropy plot of the spike gene receptor-binding domain (RBD), where low entropy values indicate highly conserved regions selected for sgRNA and primer design. (B) Phylogenetic tree illustrating the predominance of Alpha, Delta, and Omicron variants circulating in the Iranian population. (C) Genome-wide mutation frequency plot showing increased variability within the spike gene region, reflecting strong selective pressure.

The entropy and mutation analyses revealed that the receptor-binding domain (RBD) of the spike gene contains relatively conserved regions appropriate for sgRNA targeting, despite neighboring loci exhibiting elevated sequence variability. Phylogenetic reconstruction using Nextclade further confirmed that Alpha, Delta, and Omicron variants were the most prevalent clades circulating among the Iranian population during the studied period. The inclusion of these 3 major SARS-CoV-2 variants in the bioinformatic design phase highlights the broad detection capability of the developed CRISPR-Cas13a assay, ensuring robustness against key viral mutations.

These combined analyses provided a comprehensive overview of the evolutionary dynamics of SARS-CoV-2 in Iran and guided the selection of conserved regions within the spike gene for sgRNA design. To further illustrate these findings, Figure 2 presents the phylogenetic and mutational landscape of major variants, highlighting key lineage distributions, mutation hotspots, and conserved RBD residues chosen for CRISPR-based detection.

Figure 2.

Phylogenetic and mutational landscape of SARS-CoV-2 variants in Iran. (A) Phylogenetic clustering shows the predominance of Alpha, Delta, and Omicron lineages. (B) Genome-wide distribution of mutations highlights the spike gene as the most variable region. (C) Amino acid substitutions within the RBD emphasize conserved sites selected for sgRNA design.

Data were obtained from the GISAID database and analyzed using Nextclade. These findings illustrate the rapid viral evolution and the need for diagnostics that remain robust against emerging variants.

In order to complement the phylogenetic analysis and provide a clearer context for diagnostic design, we also examined the major mutation sites across prevalent SARS-CoV-2 variants. While the phylogenetic tree illustrates lineage distribution, the rapid emergence of recombinant variants such as XBB and BA.2.75 highlights the dynamic nature of viral evolution. A summary of these variant-specific mutations is provided in Table 2, offering insight into their potential impact on diagnostic sensitivity and the importance of targeting conserved regions within the spike gene. ¹⁸

Table 2.

Key Spike Mutations in Major SARS-CoV-2 Variants and Their Potential Diagnostic Relevance.

Variant category	Variant/Sublineage	Key spike mutations	Functional/diagnostic relevance	Reference
VOI	XBB.1.5 (incl. ABB.1.5)	F486P, R346T, N460K, D614G	Increased transmissibility, immune escape	PMID: 37214653
VOI	XBB.1.1.6	F486P, T478K, S486P, D614G	Antibody evasion, altered RBD affinity	PMID: 37823372
VUM	BA.2.75 (and 2A.2.75)	G446S, N460K, D339H, R493Q	Strong immune evasion, relevant to diagnostic primer binding	PMID: 36198205
VUM	XBB (excl. XBB.1.5, incl. XBB.1.9.1, XBB.1.9.2, XBB.2.3)	R346T, K444T, V445P, F486S	Recombination lineage; immune escape mutations in RBD	PMID: 37697984
VOC	Omicron (B.1.1.529 + BA lineages)	N501Y, E484A, K417N, G446S, Q498R, Y505H, Δ69–70	High immune escape, may affect sgRNA design	PMID: 36894768
VOC	Delta (B.1.617.2 + AY)	L452R, T478K, P681R, D614G	Increased infectivity, potential reduced diagnostic sensitivity	PMID: 34990727
VOC	Alpha (B.1.1.7 + Q)	N501Y, Δ69–70, P681H, D614G	Enhanced ACE2 binding, S-gene target failure in diagnostics	PMID: 33688656

Selection and Evaluation of Target Sequences

Among the numerous candidate sequences initially identified, those located within highly conserved genomic regions—based on alignment outputs from BioEdit and MEGA X—were given priority. These regions were defined through multiple sequence alignments of 470 SARS-CoV-2 genomes isolated from Iranian patients, along with comparative analyses against SARS-CoV and MERS-CoV sequences. Visual inspection of the alignments, along with E-values obtained from BLAST searches, further supported the selection process.

In addition to sequence conservation, several critical criteria were applied to target selection. These included the presence of a protospacer flanking site (PFS), optimal GC content, and sufficiently long conserved flanking regions, all of which are essential for robust primer design.

Although the ideal sgRNA target length generally falls between 22 and 28 nucleotides—with 28 nucleotides considered theoretically optimal ¹⁹ —BLAST analyses revealed significant sequence homology between 28-nt SARS-CoV-2 candidates and SARS-CoV counterparts, as indicated by low E-values.

Therefore, candidate sequences of 22, 24, 26, and 28 nucleotides in length were systematically evaluated. Among these, the 22-nt targets exhibited the highest specificity, demonstrating minimal off-target homology with SARS-CoV. ²⁰

Based on sgRNA performance scores and predicted genome-wide targeting efficiency, the candidate sequences summarized in Table 1 demonstrated over 98% coverage across SARS-CoV-2 isolates. These regions were further validated using the UCSC Genome Browser and confirmed to reside within highly conserved loci among Iranian samples. Accordingly, they were selected as optimal sgRNA targets and subsequently used as templates for RPA primer design.

Final Candidate Regions for sgRNA Design

Building upon the aforementioned conservation and specificity assessments, the sgRNA candidate sequences listed in Table 3 were selected as final targets. These sequences were subsequently evaluated using RNA secondary and tertiary structure modeling and employed for RPA primer design.

Table 3.

Summarizes the Characteristics of Each Selected sgRNA Target, Including GC Content, Genomic Position, Sequence Length, and the Number of Single Nucleotide Variants (SNVs) Observed Among SARS-CoV-2 Isolates From the Iranian Population.

N	Target site	sgRNA (crRNA)	Position	E-Value	SNV	Source
1	TGTCTATGCAGATTCATTTGTA	UACAAAUGAAUCUGCAUAGACA	22 744-22 765	0.023	0	UCSC
2	GATTGCTGATTATAATTATAAA	UUUAUAAUUAUAAUCAGCAAUC	22 813-22 834	0.023	1	UCSC
3	ATTGCTGATTATAATTATAAAT	AUUUAUAAUUAUAAUCAGCAAU	22 814-22 835	0.023	0	UCSC
4	TTGCTGATTATAATTATAAATT	AAUUUAUAAUUAUAAUCAGCAA	22 815-22 836	0.023	0	UCSC
5	TGCTGATTATAATTATAAATTA	UAAUUUAUAAUUAUAAUCAGCA	22 816-22 837	0.023	0	UCSC
6	ACCATACAGAGTAGTAGTAC	GUACUACUACUCUGUAUGGU	23 080-23 101	0.36	0	Articles
7	ACTGAAATCTATCAGGCCGGTA	UACCGGCCUGAUAGAUUUCAGU	23 003-23 024	0.023	1	BioEdi/ Entropy plot
8	TAATGGTGTTGAAGGTTTTAAT	AUUAAAACCUUCAACACCAUUA	23 035-23 056	0.023	3	BioEdi/ Entropy plot
9	TGTTACTTTCCTTTACAATCAT	AUGAUUGUAAAGGAAAGUAACA	-23057 23078	0.023	4	BioEdi/Entropy plot

Primer Design for Recombinase Polymerase Amplification (RPA)

According to previous literature and established RPA design guidelines, the optimal length for primers ranges between 25 and 35 nucleotides, while the recommended amplicon size lies between 80 and 140 base pairs. The ideal melting temperature (Tm) for RPA primers is typically within the range of 54°C to 67°C. Based on these criteria, Table 4 presents the selected forward and reverse primers used for the amplification of target regions via RPA.

Table 4.

Summary of Forward and Reverse Primers Designed for RPA Amplification.

Primer	Position	Primer length	GC (%)	Tm
F1	22 705	25	36	56.4
F2	22 711	25	32	54.6
F3	22 715	25	32	55.3
F4	22 756	25	28	53.6
F5	22 759	25	32	53.6
F6	22 765	25	28	51.5
F7	22 773	25	32	56.1
F8	22 831	25	52	65.3
F9	22 959	25	28	53.1
F10	23 050	25	35	55.3
R1	22 775	25	32	55.3
R2	22 794	25	28	52.1
R3	22 828	25	52	66.6
R4	22 832	25	52	65.4
R5	22 884	25	44	60.5
R6	22 890	25	40	57.6
R7	22 934	25	32	55.7
R8	23 054	25	32	55.9
R9	23 057	25	24	54.5
R10	23 147	25	32	56.7

The table lists selected primers designed based on conserved genomic loci flanking sgRNA target regions. Each entry includes primer position relative to the SARS-CoV-2 reference genome (Wuhan-Hu-1, NC_045512), primer length, GC content, melting temperature (Tm), dimer formation free energy (ΔG), number of hydrogen bonds at the 3′ end, and the presence or absence of potential hairpin loops. All primers met the recommended criteria for RPA, including optimal length (25-35 nt), Tm (54°C-67°C), and absence of significant secondary structures.

Multiple loci within the RBD region were evaluated (Table 4), and primer pairs were selected based on the following stringent criteria:

• 1- Absence of secondary structures such as hairpins and internal loops,

• 2- Minimal primer-dimer formation, especially at the 3′ termini,

• 3- Closely matched melting temperatures (Tm) between forward and reverse primers,

• 4- Appropriate GC content and amplicon length suitable for efficient RPA performance.

All selected primers underwent rigorous in silico validation and were deemed optimal for amplifying conserved target regions, enabling high-sensitivity CRISPR-Cas13a-based detection in downstream applications. In addition to length, GC content, and Tm, all primers were evaluated for potential secondary structures, 3′-end stability, and dimer formation; primers meeting all criteria were selected. Detailed thermodynamic and structural metrics are provided in Supplemental Table S1.

Final Primer Candidates for RPA

To ensure high target specificity, all primer candidates were initially screened using BLAST analyses against the SARS-CoV-2 genome and related coronaviruses. Their amplification efficiency and compatibility with the selected sgRNA binding regions were subsequently assessed. The finalized primer sets, as outlined in Table 5, were selected based on alignment accuracy, thermodynamic stability (including melting temperature and GC content), and the absence of secondary structures or primer-dimer formations. These primers were optimized to effectively amplify conserved segments within the receptor-binding domain (RBD) of the spike (S) gene for downstream CRISPR-Cas13a detection.

Table 5.

Final RPA Primer Sets Corresponding to the Selected Target Regions Within the Spike Protein’s Receptor-Binding Domain (RBD).

Target sequences	Primer	F/R start position Number	Length	GC (%)	Tm
1 (22 744-22 765)	F3-R4	22 715/22 832	142 nt	35.2	78.3
	F3-R1	22 715/22 775	85 nt	30.6	74
2 (22 813-22 833) 3 (22 814-22 834) 4 (22 815-22 835) 5 (22 816-22 836) 7 (23 003-23 024)	F9-R8	22 959/23 054	120 nt	35	77.5
	F9-R9	22 959/23 057	123 nt	34.1	77.3
6 (23 080-23 101) 8 (23 035-23 056) 9 (23 057-23 079)	F10-R10	23 050/23 147	122 nt	35.2	77.7

Each primer pair was rigorously evaluated for target alignment accuracy, thermodynamic stability, and compatibility with the adjacent sgRNA binding sites.

RNA Secondary Structure Prediction Using RNAfold

The secondary structures of the candidate amplicons were predicted using the RNAfold web server, which enabled thermodynamic modeling of sgRNA target regions amplified by the designed primers (amplicon lengths: 80-140 nucleotides). The key metric for evaluating RNA structural stability was Gibbs free energy (ΔG), with less negative ΔG values indicating reduced secondary structure stability and, therefore, greater accessibility for sgRNA binding—a crucial factor for effective CRISPR-Cas13a–mediated recognition and cleavage. ²¹

As shown in Figure 3, the ΔG values across the predicted amplicons were relatively close, suggesting similar thermodynamic profiles and structural accessibility. Given this, secondary structure analysis alone was insufficient for final candidate selection. To achieve more precise discrimination, additional analyses—including RNA–RNA interaction modeling and 3D structural prediction—were conducted in subsequent steps.

Figure 3.

Predicted secondary structures of candidate amplicons using RNAfold. The thermodynamic stability of the RNA amplicons was assessed using Gibbs free energy (ΔG) values. Amplicons with less negative ΔG values exhibit reduced structural stability, indicating greater accessibility for sgRNA binding.

Tertiary Structure and RNA-RNA Interaction Analysis Using RNAup

In this section, the interactions between the designed sgRNAs and their corresponding RNA target sequences were analyzed using the RNAup tool. The results are depicted in Figure 4. A lower (ie, more negative) interaction free energy (ΔG) signifies stronger binding affinity, suggesting a more thermodynamically stable sgRNA–target duplex. Therefore, sgRNAs associated with lower ΔG values were considered the most suitable candidates for downstream CRISPR-Cas13a-mediated detection. ²²

Figure 4.

Shows the results of interaction energy analysis. Red lines represent the interactive free energy of binding, while black plots indicate the energy required to unfold structured regions of the target RNA. In plots A, B, E, F, G, and H, the presence of a single binding site with a minimum free energy peak suggests specific and high-affinity binding between the sgRNA and its corresponding target.

In addition to ΔG magnitude, the length of the binding region was taken into account, as longer hybridization zones often contribute to enhanced interaction stability. All predicted RNA-RNA duplexes were also evaluated in the context of local secondary structures to ensure the compatibility and accessibility of the binding sites, thus supporting efficient sgRNA engagement. ²³

The strong interaction energies observed for candidate sgRNA–amplicon pairs indicate robust binding, suggesting that these complexes are likely to perform reliably in routine laboratory assays and facilitate practical implementation of the diagnostic workflow. ¹⁰

The selected sgRNA–amplicon pairs demonstrated favorable thermodynamic characteristics, including accessible secondary structures and strong interaction profiles based on RNA–RNA duplex modeling. As illustrated in Figure 4, interaction energy plots—particularly for candidate pairs E and F—exhibited the lowest free energy values, indicating highly stable and specific binding events.

Table 6 summarizes the predicted thermodynamic parameters of sgRNA–target interactions as determined by RNAup. These parameters include binding free energy (ΔG), interaction length, and the exact coordinates of each binding site. Collectively, these data confirm the hybridization efficiency and structural compatibility of the selected sgRNAs with their corresponding target regions, supporting their effective use in downstream CRISPR-Cas13a detection assays.

Table 6.

Thermodynamic Parameters of sgRNA–Target Interactions Predicted by RNAup. Binding free energy (ΔG), interaction length, and binding site coordinates were computed using RNAup to evaluate the hybridization potential between sgRNAs and their target amplicons.

N	Total free energy of binding (kcal/mol)	Energy from duplex formation (kcal/mol)	Opening energy for the longer sequence (kcal/mol)	Opening energy for the shorter sequence (kcal/mol)
A	−24.53	−33.91	8.90	0.48
B	−24.41	−33.91	9.01	0.48
C	−4.07	−5.14	1.00	0.08
D	−4.8	−5.14	0.99	0.08
E	−30.62	−40.41	6.33	3.46
F	−30.62	−40.41	6.33	3.46
G	−24.90	−34.70	8.24	1.56
H	−7.13	−15.20	8.07	0.00
I	−7.29	−11.81	4.43	0.10

Three-Dimensional Structure Prediction and Docking Assessment

At this stage, the 3-dimensional structures of the selected sgRNA–amplicon complexes (models A, B, E, F, G, and H) were analyzed using the HNADOCK web server. For each complex, 10 potential RNA–RNA interaction conformations were generated, each accompanied by a molecular docking score (Figure 5).

Figure 5.

Predicted 3D interaction models and molecular docking scores of sgRNA–amplicon complexes generated by HNADOCK. Docking energy values were used to evaluate the structural stability and binding affinity of each complex. The most negative score corresponds to the most thermodynamically favorable configuration.

Among the predicted configurations, the model with the lowest docking energy was considered the most thermodynamically favorable and indicative of optimal binding. ²⁴ Notably, the Amplicon 2–sgRNA 1 complex exhibited the most favorable docking score, suggesting the highest binding affinity and structural compatibility in 3-dimensional space. Such thermodynamically favorable configurations support high-efficiency target recognition and indicate potential for effective application of these CRISPR-Cas13a assays across diverse healthcare settings, including resource-limited environments. ²⁴

Discussion

Numerous studies have demonstrated that SARS-CoV-2 continues to evolve rapidly, with emerging mutations posing a potential threat to the accuracy and reliability of conventional diagnostic strategies such as RT-PCR. For instance, RT-PCR, considered the gold standard for SARS-CoV-2 detection, has limited accuracy in reliably identifying newly emerging SARS-CoV-2 variants, such as the Omicron VoC. In response to these limitations, we propose the use of the CRISPR-Cas13a system in combination with recombinase polymerase amplification (RPA) as a rapid and sensitive diagnostic approach for the detection of SARS-CoV-2. ²⁵

We hypothesize that the CRISPR-Cas13a platform can effectively detect viral RNA post-infection. However, given the limited length of single-guide RNA (sgRNA), precise selection of the target site is critical to ensure accurate and efficient recognition. Without careful target design, the diagnostic performance of the system may be significantly reduced.

Cross-contamination is a significant challenge in molecular diagnostics, particularly in low-resource settings, potentially leading to false positives and compromising assay reliability. ²⁶ Mitigation of cross-contamination in conventional PCR laboratories usually requires spatial segregation, with dedicated zones for reagent preparation, sample processing, nucleic acid amplification, and post-amplification analysis. This significantly increases infrastructural costs, including laminar flow hoods, real-time PCR machines, and highly specialized personnel. Alternative decontamination strategies, such as ultraviolet irradiation and enzymatic hydrolysis, often add procedural complexity.

CRISPR-based systems, such as SHERLOCK and DETECTR, minimize the risk of cross-contamination through single-tube reactions that integrate isothermal amplification, target recognition, and collateral cleavage. This all-in-one-tube architecture inherently reduces aerosol formation, decreases sample carryover, and enables point-of-care testing, enhancing cost efficiency and workflow simplicity.

To identify highly specific regions within the receptor-binding domain (RBD) of the spike (S) gene, we conducted comprehensive sequence alignments and structural comparisons between SARS-CoV-2 and its closely related coronaviruses, SARS-CoV, and MERS-CoV. A total of 470 complete genome sequences of Iranian SARS-CoV-2 isolates with reported clinical symptoms were retrieved from the GISAID database and included if they were full-length and of high coverage; sequences with missing data or low coverage were excluded. ²⁷ Conserved regions with the lowest Shannon entropy values and minimal similarity to related coronaviruses were selected as ideal targets for sgRNA design. Multiple candidate sgRNAs were then evaluated based on GC content, presence of protospacer flanking sites, and avoidance of homopolymeric sequences, ensuring high specificity and optimal binding efficiency.

These regions were prioritized because they are critical for viral entry and represent key targets for diagnostic assays. Phylogenetic analysis using the Nextclade web server confirmed that Alpha, Delta, and Omicron were the most prevalent variants among the Iranian population. As highlighted in the Results section, the bioinformatic design phase incorporated the major SARS-CoV-2 variants (Alpha, Delta, and Omicron), ensuring that the developed CRISPR-Cas13a assay maintains broad detection capability against key viral mutations. Furthermore, the spike gene exhibited the highest mutation rate across the genome, emphasizing the importance of selecting conserved subregions within the RBD for accurate CRISPR-based diagnostics. ²⁸ The continuous evolution of SARS-CoV-2, including sublineages BA.4 and BA.5, can reduce the sensitivity of molecular assays such as RT-qPCR. ²⁹ By targeting conserved regions, our bioinformatics-based approach aims to maximize detection across all existing and emerging variants. Nevertheless, recent evidence highlights the importance of confirming sgRNA performance against newly emerging variants. ³⁰ Therefore, experimental validation using qRT-PCR or viral culture assays will be required in future studies to ensure the robustness and real-world applicability of the designed sgRNAs.

Following this, multiple sgRNA candidates were designed to specifically target these conserved sequences. BLAST analysis revealed low E-values for some candidates, indicating high sequence similarity with SARS-CoV and highlighting the necessity of optimizing sgRNA specificity. Previous studies have shown that the length of sgRNA plays a critical role in Cas13a activity, with guides between 19 and 28 nt providing optimal RNA cleavage activity. ³¹ Bioinformatic filters were applied to enhance specificity, including a preference for sgRNAs with a 5′ terminal cytosine or guanine, a GC content between 25% and 65%, and the absence of homopolymeric regions. Each sgRNA was composed of a conserved stem-loop scaffold and a variable protospacer sequence. ³² The thermodynamic properties of each sgRNA–amplicon pair were assessed using RNAfold and RNAup. Six sgRNA–amplicon pairs (A, B, E, F, G, H) demonstrated favorable interaction profiles with strongly negative ΔG values, indicating strong and specific hybridization potential. ³³ These results provide insight into the structural principles governing sgRNA–target recognition, which can guide rational sgRNA design for other viral or genomic targets. Moreover, RNAfold and RNAup proved to be reliable predictors of sgRNA–RNA interactions, supporting their utility in computational CRISPR assay development. ³⁴

The reliance on an exclusively in silico approach is both a strength and a limitation of the present study. Bioinformatics methods enable rapid and cost-effective assessment of sgRNA candidates, providing a valuable framework for diagnostic design, particularly in resource-limited settings or at the onset of emerging infectious threats. However, computational predictions cannot fully account for biological complexities such as RNA secondary structures, off-target effects in cellular environments, or assay performance under clinical conditions. Planned next steps include in vitro validation using RT-RPA and qRT-PCR assays to assess sensitivity, specificity, and robustness of the selected sgRNAs under real diagnostic conditions. This stepwise progression, beginning with computational screening and advancing toward experimental confirmation, has been widely recognized as an effective strategy in CRISPR-based diagnostic development. ³⁵

The CRISPR-Cas13a system has demonstrated significant potential for the detection of various RNA viruses, including SARS-CoV-2, hepatitis D virus (HDV), and plant RNA viruses. ³⁵ Its high specificity and sensitivity make it a promising tool for rapid diagnostics. By leveraging computational tools like RNAfold and RNAup, we have identified optimal sgRNA–amplicon pairs with favorable thermodynamic properties, enhancing the efficiency of the CRISPR-Cas13a system in detecting SARS-CoV-2. These findings underscore the versatility of CRISPR-Cas13a-based diagnostics and its applicability to a broad range of RNA viruses.

Limitations

This study was conducted entirely in silico, focusing on the bioinformatic design and evaluation of CRISPR-Cas13a sgRNAs targeting the SARS-CoV-2 spike region. Therefore, no sample size calculation or power analysis was performed, which represents a limitation for extrapolating these findings to real-world diagnostic settings.

Although the computational approach provides valuable insights into the feasibility and specificity of the designed sgRNAs, the absence of laboratory-based validation represents a key limitation. Future studies should therefore include experimental confirmation through in vitro assays such as RT-RPA, qRT-PCR, or viral culture-based detection tests to substantiate these findings and ensure broad applicability to emerging variants. ³⁶ Similar computational-only studies have been published as an initial step prior to experimental work, ³⁷ highlighting the role of bioinformatics in laying the groundwork for subsequent wet-lab investigations. Therefore, the results presented here should be interpreted as preliminary, forming the basis for future experimental validation.

Conclusion

The 6 selected Amplicon–sgRNA complexes demonstrated the highest guide activity among all tested candidates. To gain deeper insights into their molecular interactions, comprehensive docking simulations were conducted. Among these, the Amplicon 2–sgRNA 1 complex exhibited the lowest molecular docking energy, indicating the most stable 3-dimensional conformation and identifying it as the most promising candidate for diagnostic application.

Upon delivery into host cells, the Cas13a–sgRNA1 complex becomes activated upon recognition of the target viral RNA, enabling both identification and subsequent degradation of viral genomic RNAs and mRNAs. This cleavage mechanism effectively disrupts viral RNA replication, protein translation, and virion assembly, thereby impairing the viral life cycle.

Importantly, the CRISPR-Cas13a platform holds significant potential not only for the detection of SARS-CoV-2, but also for the diagnosis and possible therapeutic intervention of other RNA viruses, including SARS-CoV and MERS-CoV. ³⁸ Beyond its application in COVID-19 diagnostics, the CRISPR-Cas system represents a powerful adjunct to existing vaccination and antiviral strategies and offers a versatile platform for combating a wide range of RNA virus-associated infectious diseases. Future in vitro and clinical validation will be essential to translate these bioinformatic findings into practical diagnostic tools. This study represents the initial phase of a larger research project, and the lack of experimental verification is recognized as a limitation of this preliminary stage. The single-tube design of CRISPR-Cas13a assays also minimizes cross-contamination, enhancing reliability in practical diagnostic settings, and future phases are planned to provide the necessary empirical evidence.