key: cord-0725354-z5eyih5o
authors: Xiao, Yan; Li, Zhen; Wang, Xinming; Wang, Yingying; Wang, Ying; Wang, Geng; Ren, Lili; Li, Jianguo
title: Comparison of three TaqMan real-time reverse transcription-pcr assays in detecting SARS-CoV-2
date: 2020-12-01
journal: J Virol Methods
DOI: 10.1016/j.jviromet.2020.114030
sha: c8e3d026db0665d150bb2bf7920a5167a63ac389
doc_id: 725354
cord_uid: z5eyih5o

Quick and accurate detection of SARS-CoV-2 is critical for COVID-19 control. Dozens of real-time reverse transcription PCR (qRT-PCR) assays have been developed to meet the urgent need of COVID-19 control. However, methodological comparisons among the developed qRT-PCR assays are limited. In the present study, we evaluated the sensitivity, specificity, amplification efficiency, and linear detection ranges of three qRT-PCR assays, including the assays developed by our group (IPBCAMS), and the assays recommended by WHO and China CDC (CCDC). The three qRT-PCR assays exhibited similar sensitivities, with the limit of detection (LoD) at about 10 copies per reaction (except the ORF 1b gene assay in CCDC assays with a LoD at about 100 copies per reaction). No cross reaction with other respiratory viruses were observed in all of the three qRT-PCR assays. Wide linear detection ranges from 10(6) to 10(1) copies per reaction and acceptable reproducibility were obtained. By using 25 clinical specimens, the N gene assay of IPBCAMS assays and CCDC assays performed better (with detection rates of 92 % and 100 %, respectively) than that of the WHO assays (with a detection rate of 60 %), and the ORF 1b gene assay in IPBCAMS assays performed better (with a detection rate of 64 %) than those of the WHO assays and the CCDC assays (with detection rates of 48 % and 20 %, respectively). In conclusion, the N gene assays of CCDC assays and IPBCAMS assays and the ORF 1b gene assay of IPBCAMS assays were recommended for qRT-PCR screening of SARS-CoV-2.

Since the first detection in late 2019, severe respiratory syndrome CoV-2 (SARS-CoV-2) caused Corona Virus Infectious Disease in 2019 has widely spread in the world. By April 11, 2020, more than 1.7 million patients infected by SARS-CoV-2 has been reported from 185 countries (Dong, Du, and Gardner, 2020) .

Given the quick increase in confirmed cases and asymptomatic infections, there are increasing demands in diagnostic tools for quick and accurate detection of the virus (Corman et al., 2020; Phan, 2020) . Several real-time reverse transcription-Polymerase Chain Reaction (qRT-PCR) for the detection of SARS-COV-2 has been developed to meet the demands, including the assays by this group (IPBCAMS [Institute of Pathogen Biology, Chinese Academy of Medical Sciences] assays), and the assays by WHO (WHO assays), and the assays by China CDC (CCDC assays).

Because SARS-CoV-2 usually infected the lower respiratory tract, it is not easy to detect the viral nucleic acids from throat swabs with relatively lower viral load (Zou et al., 2020) . Thus, qRT-PCR assays with higher sensitivity and better performance in the detection of SARS-CoV-2 is recommended in aiding the diagnosis of COVID-19 (Corman et al., 2020) . However, most of the current available qRT-PCR assays were developed for emergency, a comprehensive methodological comparison among these assays remains unfulfilled.

To comprehensively compare the performance of currently available qRT-PCR assays for detection of SARS-CoV-2, we evaluated the sensitivity, specificity, amplification efficiency, and linear detection ranges among IPBCAMS assays, WHO assays and CCDC assays.

Clinical specimens (throat swabs and sputum) suspected of COVID-19 infection were collected from Jin Yin-Tan hospital. Nucleic acids were extracted from a volume of 200 μl clinical specimens by using NucliSens easyMag apparatus (bioMe´rieux, MarcyL'Etoile, France) according to the manufacturer's instructions. A volume of 50 J o u r n a l P r e -p r o o f μl total nucleic acid eluate for each specimen was recovered and transferred into a nuclease-free vial and either tested immediately or stored at -80°C. Clinical specimens from healthy volunteers were applied as negative control. A human housekeeping gene (GAPDH) was employed as internal control. This study was approved by the Ethical Review Board of Institute of Pathogen Biology, Chinese Academy of Medical Sciences & Peking Union Medical College, and the Institutional Review Board of Jin Yin-Tan Hospital.

Sequences of primers and probes for the IPBCAMS assays were recently developed (Yiwei et al., 2020) , which were designed to exactly matched the genome of SARS-CoV-2 and had low sequence identity to other coronaviruses (SARS-CoV, human CoV 229E/NL63/HKU1/OC43, and Bat SARS-like CoV). The design of the primers and probes followed several principles, including: primer length: 18-25bp; probe length:20-30 bp; melting temperature (Tm) of primer: 55-60 °C, Tm of probe 60-65 °C. Sequences of primers and probes for the WHO assays were obtained from the website of WHO (https://www.who.int/docs/default-source/coronaviruse/protocol-v2-1.pdf?sfvrsn=a9ef618c_2), and those for the CCDC assays were obtained from the website of China CDC (http://www.chinacdc.cn/jkzt/crb/zl/szkb_11803/jszl_11815/202003/W020200309540 843062947.pdf) ( Table 1) . Primers and probes were synthesized by standard phosphoramidite chemistry techniques at Qingke biotechnology Co. ltd (Beijing, China). TaqMan probes were labeled with the molecule 6-carboxy-fluroscein (FAM) at the 5' end, and with the Blackhole Quencher 1 (BHQ1) at the 3' end. Optimal concentrations of the primers and probes were determined by cross-titration of serial two-fold dilutions of each primer/probe against a constant amount of purified RNA of SARS-CoV-2. No amplification signal was obtained in all of the three qRT-PCR assays with nucleic acids extracted from clinical specimens of healthy volunteers as template.

The qRT-PCR assays were performed by using TaqMan 

Limit of detection (LoD) of the three qRT-PCR assays were determined through the Probit Regression analysis. The template RNA was diluted from 100 copies per reaction through 50, 20, 10, 5 to 1 copy per reaction. Ten replicates of each dilution were applied for LoD determination. The LoD of a qRT-PCR assay was defined as the lowest detectable dilution of viral RNA transcript with a 95% probability in the Probit Regression analysis.

Linear detection ranges of the three qRT-PCR assays were determined through 10-fold serial dilutions of the RNA transcripts as template. Fitting curve between the Ct values and quantities of the RNA transcript were applied for evaluation of the detection linearity of the assays. A good linearity was defined with a correlation coefficient (r 2 ) higher than 0.99 in the fitting curve. Efficiency of the three qRT-PCR assays were evaluated by the slope of the fitting curve, which was defined as 10 (−1/slope) -1.

Reproducibility of the three qRT-PCR assays were assessed by the coefficient of variation of the Ct values of the 10-fold serial diluted RNA transcripts in the intra-J o u r n a l P r e -p r o o f and the inter-assay. Triple replicates of each dilution were applied in the intra-assay.

The inter-assay consisted of triple replicates of the intra-assay. The coefficient of variation was calculated by the standard deviation of the Ct values of an RNA dilution divided by the mean Ct value of the same RNA dilution.

Nucleic acids of common respiratory viruses, extracted by using a NucliSens easyMag apparatus (bioMe´rieux, MarcyL'Etoile, France) according to the manufacturer's instructions, were applied as templates for evaluation of potential cross-reactions of the three qRT-PCR assays, including human coronaviruses (OC43, NL63, 229E, and HKU1), Influenza viruses (A or B), respiratory syncytial virus, parainfluenza virus (1-4), human metapneumovirus, rhinovirus, adenovirus, and bocavirus.

A serial dilution panel of the RNA transcript was tested to determine the LoD of the three qRT-PCR assays, defined as the minimum concentration with detection of 95% by Probit regression analysis. The 95% detection limit of the N gene assay were 9.7 copies per reaction (95% CI 7.4-15.2), 6.6 copies per reaction (95% CI 4.9-13.1), and 10.5 copies per reaction (95% CI 7.9-17.1) for the IPBCAMS assay, the CCDC assay, and the WHO assay, respectively. The 95% detection limit of the ORF 1b gene assay were 27.8 copies per reaction (95% CI 20.7-48.9), 33.6 copies per reaction (95% CI 27.1-55.8), and 23.1 copies per reaction (95% CI 17.6-37.0) for the IPBCAMS assay, the CCDC assay, and the WHO assay, respectively.

The linear detection ranges of the three qRT-PCR assays were determined by using a ten-fold dilution of the RNA transcript as template. Strong linear correlations (Table   2 ) were observed between the Ct values and quantity of RNA transcripts with r 2 =0.9926, 0.9987 in the N gene assay, and r 2 =0.9953, 0.9941 in the ORF 1b assay of IPBCAMS assays and CCDC assays, respectively. Good linear correlations (Table 2) were observed in WHO assays, with r 2 =0.9750 and 0.9897 for the N gene assay and the ORF 1b assay, respectively. These results suggested that all of the three qRT-PCR assays exhibited linear detection ranges from 10 6 to 10 1 copies per reaction, while the WHO assays showed lower coefficient of linear correlation.

The reproducibility of the three qRT-PCR assays was assessed by measuring coefficient of variation (CV) of the Ct values in the intra-and inter-assay (Table 3) .

For the N gene assay, the CVs of mean Ct values from 10 6 to 10 1 copies of RNA transcript per reaction were 0.20%-1.33%, 0.46%-5.09%, 0.27%-1.97% in intra-assay, and 1.06%-2.45%, 0.96%-7.59%, 1.00%-5.51% in inter-assay of IPBCAMS assay, WHO assay, and CCDC assay, respectively. While the N gene assay in WHO assays exhibited relative high CVs with 0.46%-5.09% and 0.96%-7.59% in the intra-and inter-assay, respectively. For the ORF 1b gene assay, the CVs of mean Ct values were 0.26%-4.45%, 0.29%-1.76%, 0.71%-6.52% in intra-assay, and 2.17%-5.12%, 0.30-1.57%, 2.63%-4.34% in inter-assay of IPBCAMS assays, WHO assays, and CCDC assays, respectively.

Because co-infections of respiratory viruses are common, we prepared a mixture of the RNA transcript and a pooled total nucleic acid extract from respiratory specimens (RNA transcript + other extract, v: v=1:1) as template, to evaluate the effect of coexisted viral nucleic acids on the performance of the assays. The co-existed other viral nucleic acids increased the Ct values of SARS-CoV-2 in most of the qRT-PCR assays, except for the ORF 1b gene assay of the CCDC assays (Table 2) . Increased amplification efficiency of SARS-CoV-2 with the co-existed other viral nucleic acids, were observed in all the three qRT-PCR assays ( Table 2) .

To evaluate the potential cross-reactions with other human respiratory viruses, the three qRT-PCR assays were examined by using human respiratory samples as templates, which were positive for human coronaviruses (OC43, NL63, 229E, or HKU1), or Influenza viruses (A or B), or respiratory syncytial virus, or parainfluenza virus (1-4), or human metapneumovirus, or rhinovirus, or adenovirus, or bocavirus.

No cross reaction was observed in all of the three qRT-PCR assays (data not shown), suggesting high specificity of the three qRT-PCR assays in detecting SARS-CoV-2.

The three qRT-PCR assays were evaluated with 25 clinical specimens (including 13 throat swabs and 12 sputum) from 25 suspected COVID-19 patients (Table 4) . SARS-CoV-2 was detected from 92% (23/25), 60% (15/25), 100% (25/25) by the N gene assay, and from 64% (16/25), 48% (12/25), 20% (5/25) of all enrolled clinical specimens by the ORF 1b gene assay in IPBCAMS assays, WHO assays, CCDC assays, respectively (Table 4) . With respect to the sputum, SARS-CoV-2 was detected from 100% (12/12), 75% (8/12), 100% (12/12) of specimens by the N gene assay, and from 100% (12/12), 75% (8/12), 41.7% (5/12) of specimens by the ORF 1b gene J o u r n a l P r e -p r o o f assay in in IPBCAMS assays, WHO assays, CCDC assays, respectively. About the throat swabs, SARS-CoV-2 was detected from 84.6% (11/13), 53.8% (7/13), 100%

(12/12) of specimens by the N gene assay, and from 30.8% (4/13), 30.8% (4/13), 0% (0/13) of specimens by the ORF 1b gene assay in in IPBCAMS assays, WHO assays, CCDC assays, respectively. These results demonstrated that the N gene assay performed better than the corresponding ORF 1b gene assay of all the three qRT-PCR assays, the N gene assay in CCDC assays and ORF 1b gene assay in IPBCAMS assays performed better than the other assays.

Rapid and accurate detection of SARS-CoV-2 represents a fast-growing global demand, which could be met by qRT-PCR. However, the current available qRT-PCR assays for SARS-CoV-2 vary in performance, including sensitivity, specificity, reproducibility, linear detection ranges, etc. Moreover, because the viral load of SARS-CoV-2 in upper respiratory tract is relatively low, reliable qRT-PCR assays for the detection of SARS-CoV-2 are required for accurate diagnosis of COVID-19. We thus compared the performance of three currently wide-applied qRT-PCR assays in the detection of SARS-CoV-2.

Sensitivity is the primary demand in the detection of respiratory viruses (Huang et al., 2018) . The three qRT-PCR assays provide LoDs of 6.6-33.6 genomic copies per reaction with a detection range from 10 6 -10 1 genomic copies per reaction. These results suggested that most of the three qRT-PCR assays provide high sensitivity and wide linear detection range in detecting SARS-CoV-2, except a relative lower sensitivity observed in the ORF 1b gene assay of CCDC assays.

Specificity is also essential in the detection of SARS-CoV-2, because of common coinfections with other respiratory viruses and high host DNA background in throat swabs (Kim et al., 2013; Touzard-Romo, Tape, and Lonks, 2020; Wu et al., 2020) . We evaluated the specificity of the three qRT-PCR assays with respiratory specimens positive for other common respiratory viruses. No cross reaction was observed, demonstrating high specificity of the three qRT-PCR assays in detection of SARS-CoV-2.

We next evaluated the reproducibility of the three qRT-PCR assays by measuring coefficient of variation (CV) of mean Ct values in intra-and inter-assay (Feng et al., 2018) . The N gene assay in IPBCAMS assays and ORF 1b gene assay in WHO assays J o u r n a l P r e -p r o o f exhibited a relative better reproducibility with lower intra-and inter-assay CVs, which were not affected by the co-existed nucleic acids of other respiratory viruses.

Efficiency is another key parameter of qRT-PCR, reflecting the binding efficiency of primers & probe to template and the amplification efficiency of the PCR system (Resa et al., 2014) . Most of the qRT-PCR assays provided good efficiency, except an abnormal efficiency of 121.83% observed in the ORF 1b gene assay of WHO assays.

An exceptionally high efficiency indicates an increased risk of false positive (Bilgrau et al., 2016) . The co-existed nucleic acids of other respiratory viruses increased the efficiency of all the three qRT-PCR assays, suggesting potential increased risk of cross-reactions between the primers & probe and background nucleic acids.

We finally evaluate the performance of the three qRT-PCR assays with clinical specimens from suspected SARS-CoV-2 infected patients .

Possibly because of the lower viral load in upper respiratory tract (Zou et al., 2020) , the detection rate of SARS-CoV-2 was lower in throat swabs than in sputum by all of the three assays. Meanwhile, the N gene assay performed better than the corresponding ORF 1b gene assay in all of the three qRT-PCR assays. For the N gene assay, IPBCAMS assays and CCDC assays performed better than WHO assays, both of which could detect SARS-CoV-2 from more than 90% of the suspected specimens.

For the ORF 1b gene assay, IPBCAMS assays performed better than WHO assays and CCDC assays, with a detection rate of 64%.

The results of qRT-PCR assay validations would be more precise with more clinical specimens. Thus, the results of the present study generated from 25 clinical specimens should be limited. Studies enrolled more clinical specimens covering all COVID-19 infected populations were recommended to make more precise validation of qRT-PCR assays for SARS-CoV-2.

In conclusion, we performed methodological evaluations on three widely-applied qRT-PCR assays for the detection of SARS-CoV-2. Although most of the evaluated assays exhibited good sensitivity, specificity, reproducibility and wide linear detection range, performance test with clinical specimens from suspected COVID-19 patients suggested that the N gene assay in IPBCAMS assays and CCDC assays, and the ORF 1b gene assays in IPBCAMS assays were the preferred qRT-PCR assays for accurate detection of SARS-CoV-2.

The original data will be available upon request. J o u r n a l P r e -p r o o f The coefficient of variation was calculated by the standard deviation of the Ct values of an RNA dilution divided by the mean Ct values of the same RNA dilution. 

Unaccounted uncertainty from qPCR efficiency estimates entails uncontrolled false positive rates

An interactive web-based dashboard to track COVID-19 in real time

A multiplex onetube nested real time RT-PCR assay for simultaneous detection of respiratory syncytial virus, human rhinovirus and human metapneumovirus

Multiplex PCR system for the rapid diagnosis of respiratory virus infection: systematic review and meta-analysis

Comparison of Anyplex II RV16 with the xTAG respiratory viral panel and Seeplex RV15 for detection of respiratory viruses

Novel coronavirus: From discovery to clinical diagnostics

Development of an efficient qRT-PCR assay for quality control and cellular quantification of respiratory samples

Co-infection with SARS-CoV-2 and Human Metapneumovirus

Co-infection with SARS-CoV-2 and Influenza A Virus in Patient with Pneumonia

Development of two TaqMan real-time reverse transcription-PCR assays for the detection of severe acute respiratory syndrome coronavirus-2

Clinical evaluation of a panel of multiplex quantitative real-time reverse transcription polymerase chain

We would like to thank the clinicians who contributed to sample collection and transportation. This study was funded in part by the Project from the Ministry of 

LiLi Ren: Conceptualization, data curation. Yan Xiao: Writing -Original draft preparation. Zhen Li, software, Validation. Xinming Wang, Ying Wang and Geng Wang: Investigation. Jianguo Li: Writingreview & editing, Supervision.

The authors declare that there are no conflicts of interest regarding the publication of this paper.