COV2-ID, a MIQE-compliant sub-20-min 5-plex RT-PCR assay targeting SARS-CoV-2 for the diagnosis of COVID-19

The study addresses the need for fast, sensitive, and reliable detection of SARS-CoV-2 to enable accurate diagnosis and inform public health strategies. Existing RT-qPCR tests often use lengthy protocols with real-time data acquisition and may exhibit suboptimal sensitivity, potential false negatives due to viral mutations, sampling issues, or inhibitors. The authors aim to design an MIQE-compliant assay (CoV2-ID) that targets multiple viral genes to enhance specificity and robustness, incorporates human and artificial RNA controls to assess sample quality and inhibition, and uses rapid thermal cycling and multiple cycle fluorescence detection (MCFD) to significantly shorten assay time while reducing reliance on subjective Cq values. CoV2-ID is intended to function both qualitatively and quantitatively.

The paper reviews current nucleic acid-based methods for SARS-CoV-2 detection, including RT-qPCR, isothermal amplification, and CRISPR approaches. Isothermal methods can be fast and suitable for point-of-care but require more optimization; PCR assays are simpler to design, multiplex, and generally offer greater sensitivity. Commercial RT-qPCR tests typically run in 1–1.5 hours and report sensitivities around 500 copies per reaction, though streamlined assays with ~15-copy LODs still require ~2 hours total workflow. Targeting multiple viral genes is recommended to reduce false negatives. The authors highlight concerns about mutations in primer/probe binding regions that may compromise assay sensitivity, citing evidence of mutated primer sites and early-pandemic sensitivity issues from primer mismatches. They also discuss the widespread D614G spike mutation and emphasize the need for continual assay monitoring for sequence drift. The MIQE guidelines are proposed as a framework to improve assay reliability and reporting.

Ethics: Approved by Anglia Ruskin University (HEMS-FREP 19/20/039). Samples were anonymized residual diagnostic RNAs; consent not required per UK Human Tissue Authority guidance.

Samples: Nasopharyngeal/nose-throat/throat swabs collected at Mid and South Essex NHS Foundation Trust, Broomfield Hospital. RNA extracted within 24 h, eluted in 50 µL RNase-free water with modified protocol (10 min RT at room temp in Class 1 cabinet, then 10 min at 56 °C). Stored at -80 °C. A total of 23 SARS-CoV-2-positive and 5 negative clinical RNA samples (batches A–D and E1–E5) were selected; some positives were diluted (A4 1:100; A1, C1–C7 1:30) for development/optimization. Two controls for genotyping included: cultured RNA from an Australian clinical isolate (WT) and a Twist Biosciences synthetic control (WT).

Assay design: SARS-CoV-2-specific primers/probes designed using Allele ID 7, with manual adjustments to maximize sensitivity/robustness. Targets: Nsp10, Nsp12, and N genes. Specificity checked in silico via Primer-BLAST and BLAST against SARS-CoV-1 and bat SARS-like coronavirus. JUN (intronless human gene) used as extraction/control to confirm human nucleic acid presence. Two artificial extraction/inhibition controls designed (EICAS1 and EICAS2), with EICAS2 flanked by JUN primer sites to minimize primer interference; both detected with the same probe. An LNA-based genotyping assay for the D614G (A→G) spike mutation was designed using Beacon Designer 8.2.

Oligonucleotide handling: DNA oligos resuspended at 100 µM in RNase-free water and stored at -20 °C; RNA oligos diluted 1×10⁻³ from stock and stored at -80 °C.

Clinical comparator and RT-qPCR: Hospital testing used Viasure one-tube RT-qPCR (20 µL reactions; 15 min 45 °C RT; 2 min activation; 45 cycles 95 °C 10 s / 60 °C 50 s) on Mic qPCR cyclers; positivity defined as Cq <38 for ORF1ab and N or ORF1ab alone. CoV2-ID validation used single-tube RT-qPCR (PrimeScript III, Takara; or 1Step Go, PCRBio): 5 µL reactions; typically 1 µL RNA per reaction; 5 min 50 °C RT; 1 min activation; 40 cycles of 95 °C 1 s and 60 °C 1 s. NTC and NRC controls included.

Optimization: Primer concentration optimization (0.3, 0.6, 1 µM) selected for lowest Cq; probe concentrations 0.4 vs 0.8 µM. Annealing temperatures optimized via gradient on Bio-Rad CFX with SYBR Green to confirm single melt peaks. Minimum qPCR run times established by keeping a master mix on ice and progressively reducing denaturation and polymerization times to instrument minimums (1 s each). RT time reductions tested first with PCRBio One Step (separate RT) and then with PrimeScript III (single-tube). Instruments used: Bio-Rad CFX, PCRBio Eco, Techne Prime Pro, and BMS Mic.

ddPCR: QX200 (Bio-Rad) used for absolute quantification and to establish limits of quantification (LOQ) and to support LOD estimates. Reactions used qPCR-optimized primer/probe concentrations; droplets generated and PCR run on C1000 Touch (Bio-Rad): 95 °C 10 min; 40 cycles 94 °C 30 s / 60 °C 1 min; 2 °C/s ramp; droplet reading on QX200. RT-ddPCR included a 45 °C 1 h RT step. Analysis via QuantaSoft Analysis Pro and QX Manager.

Multiple cycle fluorescence detection (MCFD): SensiFast, 5 µL volumes on Bio-Rad CFX Connect using a protocol with discrete plate reads after cycles 8, 15, 20, 25, 30, and 35. Baseline fluorescence set at cycle 8; ΔF computed at each subsequent detection cycle.

Assay composition and multiplexing: Initial four-plex combined Nsp10 (FAM), N (Texas Red), JUN (Cy5), and EICAS2 (HEX). Comparison of singleplex vs multiplex performance and EICAS1 vs EICAS2 evaluated. The assay was expanded to five-plex by adding Nsp12 as a third viral target, first sharing FAM with Nsp10, then ultimately detecting all three viral targets (Nsp10, Nsp12, N) on FAM to enhance sensitivity.

Run-time acceleration: Progressive reductions in RT duration (to as low as 1 min) and PCR step durations (1 s denaturation and 1 s annealing/polymerization), as well as narrowing temperature gaps (e.g., lowering denaturation temperature and raising annealing/polymerization temperature) to minimize cycle times. Comparative runs on faster (PCRMax/Techne) and standard (Bio-Rad CFX) instruments assessed.

Data analysis: Instrument software, Microsoft Excel, and GraphPad Prism used. Correlations between targets calculated; validation against hospital and a commercial kit (Sansure) performed.

• Specificity and assay integrity: In silico analyses confirmed primers/probes specific to SARS-CoV-2 with no cross-reactivity. No-template amplifications were absent, indicating no synthesis-associated contamination of oligonucleotides.
• PCR efficiency: All six assays showed efficient RT-qPCR with 94–103% efficiency and single melt curve peaks, indicating specific amplification.
• Quantification and detection limits: ddPCR serial dilutions showed linear quantification down to ~50 copies. Reliable quantification was established at 41 ± 12 copies (Nsp10). Translating to qPCR LOD, with Nsp10: 5 copies detected in 12/12 replicates; 2 copies in 10/12; 1 copy in 8/12. Nsp12 yielded similar results. A repeat at a predicted 2-copy level detected 24/24 reactions. Underlying data in Supplementary Table 4.
• Multiplex performance: Singleplex vs four-plex produced comparable Cqs; JUN lagged by ~2 Cq in multiplex but was corrected by increasing JUN primer concentration to 1.3 µM. EICAS1 and EICAS2 performed similarly without adverse effects on other targets.
• Clinical validation: On 28 clinical RNAs (23 positive, 5 negative), CoV2-ID results were 100% concordant with hospital testing; sensitivity 100% and specificity 100%. One hospital discordant sample (A8, ORF1ab/N discordance) was positive for both viral markers with CoV2-ID. In a subset (n=10) tested with Sansure, CoV2-ID detected ORF1ab in sample B1 missed by Sansure. Strong correlations observed: Nsp10 vs N r=0.96 (95% CI 0.89–0.98); Nsp10 vs JUN r=0.73 (0.45–0.88); N vs JUN r=0.86 (0.68–0.94).
• Inhibition assessment: EICAS-based analysis across 28 samples indicated little to no inhibition; median EICAS Cq in samples 27.07 (25.57–29.12) vs NTC median 27.08 (26.91–27.54).
• Mutation tolerance and genotyping: Primer/probe-binding site mutations (e.g., Nsp10 F primer position -9 C→T; N-gene primer positions -10, -7; probe mutations at positions 2, 3, 5) had minimal effect on performance below 65 °C, with WT and mutant primers/probes performing similarly. D614G genotyping showed all Chelmsford-area patient isolates (mid-April to June 2020) were mutant (G), while an Australian clinical sample and a Twist control were WT.
• Five-plex and same-fluorophore detection: Adding Nsp12 to create a five-plex and detecting Nsp10 and Nsp12 both on FAM reduced Cq and increased sensitivity (~80% increase suggested by ddPCR). Detecting all three viral targets (Nsp10, Nsp12, N) on FAM further enhanced sensitivity in both qPCR and ddPCR.
• Speed optimization: Reducing RT to 5 min and PCR steps to 1 s each cut run time from ~33:40 to ~20 min with no performance loss; further reducing RT to 1 min yielded Cqs comparable to 5-min RT, totaling ~16 min on suitable instruments. On Bio-Rad CFX, a fast protocol (1 min RT; 1 s/1 s cycles) reduced run time from 58 to 32 min with similar or slightly lower Cqs. Narrowing temperature gaps (lower denaturation; higher annealing/polymerization) further reduced run time to ~14 min 11 s with minimal ΔCq.
• MCFD protocol: Implementing multiple cycle fluorescence detection on CFX reduced run from ~43 min (standard qPCR) to ~22 min. Fluorescence increments at cycles 15–35 supported a 5-level diagnostic rating algorithm, enabling rapid qualitative calls without relying on subjective Cq thresholds.
• Quantification framework: ddPCR-quantified EICAS enables indirect viral load estimation and QC. Demonstrated that threshold-dependent Cq can vary by 8.7 cycles (~400-fold) for the same target in a run, whereas ddPCR copy numbers were stable (mean 1163 ± 61 copies), underscoring the benefit of EICAS/ddPCR-based quantification.

The CoV2-ID assay directly addresses the need for rapid, sensitive, and reliable SARS-CoV-2 detection while mitigating common pitfalls in RT-qPCR diagnostics. By targeting three viral genes and incorporating robust internal controls (human JUN and artificial EICAS), the assay enhances specificity and reduces the risk of false negatives due to sampling variability, inhibitors, or sequence mutations. The multiplex design and the option to detect multiple viral targets on the same fluorophore improve sensitivity, enabling consistent detection down to single-copy levels in repeated trials. Rapid thermal cycling protocols and MCFD significantly shorten time-to-result, accommodating high-throughput diagnostic workflows and point-of-care aspirations. The MCFD approach also reduces reliance on subjective Cq thresholds by employing defined fluorescence increments and a rating algorithm, clarifying qualitative interpretations and facilitating actionable decision-making. For scenarios requiring quantification, the ddPCR-calibrated EICAS provides a pathway to estimate viral load more robustly than Cq-based methods, although such quantification remains laboratory-specific and requires validation. Together, these features align with MIQE guidelines and respond to the challenges of mutation-driven assay drift, operational speed, and interpretive consistency in clinical diagnostics.

The study presents CoV2-ID, an MIQE-compliant five-plex RT-qPCR assay for SARS-CoV-2 that combines high sensitivity and specificity with rapid turnaround and robust internal controls. It reliably detects very low copy numbers, maintains performance in multiplex formats, and is resilient to certain primer/probe-binding mutations. The assay can be completed in approximately 14–20 minutes on compatible instruments and about 22 minutes using an MCFD protocol on standard platforms. The inclusion of a ddPCR-quantified artificial RNA control (EICAS) enables quality assessment and potential viral load estimation, while the MCFD algorithm simplifies qualitative diagnostics without reliance on subjective Cq thresholds. Future work could focus on broader multi-center validation, continued surveillance of viral genome variation affecting primer/probe sites, integration into point-of-care devices with extreme PCR, and standardized frameworks for inter-laboratory quantitative calibration using certified controls.

• Further validation of RNA sample quality was not possible due to diagnostic requirements; assessment of inhibitory contamination relied on the EICAS assay.
• Clinical samples were anonymized with no clinical or epidemiological data available beyond reported Cq values, limiting correlation with clinical context.
• The authors note that quantitative data derived using the EICAS/ddPCR framework will remain laboratory-specific and require separate and repeated validation in different laboratories before routine quantitative reporting.