key: cord-0856691-hdrh8i24
authors: Madala, Bindu Swapna; Reis, Andre L. M.; Deveson, Ira W.; Rawlinson, William; Mercer, Tim R.
title: Chimeric synthetic reference standards enable cross-validation of positive and negative controls in SARS-CoV-2 molecular tests
date: 2020-06-11
journal: bioRxiv
DOI: 10.1101/2020.06.09.143412
sha: c45ce82bc74421ae25a92332b0adca97ded74612
doc_id: 856691
cord_uid: hdrh8i24

DNA synthesis in vitro has enabled the rapid production of reference standards. These are used as controls, and allow measurement and improvement of the accuracy and quality of diagnostic tests. Current reference standards typically represent target genetic material, and act only as positive controls to assess test sensitivity. However, negative controls are also required to evaluate test specificity. Using a pair of chimeric A/B RNA standards, this allowed incorporation of positive and negative controls into diagnostic testing for the Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2). The chimeric standards constituted target regions for RT-PCR primer/probe sets that are joined in tandem across two separate synthetic molecules. Accordingly, a target region that is present in standard A provides a positive control, whilst being absent in standard B, thereby providing a negative control. This design enables cross-validation of positive and negative controls between the paired standards in the same reaction, with identical conditions. This enables control and test failures to be distinguished, increasing confidence in the accuracy of results. The chimeric A/B standards were assessed using the US Centers for Disease Control real-time RT-PCR protocol, and showed results congruent with other commercial controls in detecting SARS CoV-2 in patient samples. This chimeric reference standard design approach offers extensive flexibility, allowing representation of diverse genetic features and distantly related sequences, even from different organisms.

from the positive control, it could be because (i) the test failed, (ii) the reference control failed or (iii) a 48 technical issue with the testing platform. This leads to delays in diagnosis, missed diagnoses and 49 invalidation of correct test results.

The use of in vitro synthesis of RNA and DNA standards allows flexibility in control design and tailoring of 51 controls to the diagnostic test and targets. Here, we propose a new design strategy for reference 52 standards that uses matched chimeric synthetic standards in accordance with the principle of A/B testing.

In this design, all the target sequences of a molecular test are retrieved and split between groups A and 54 B, which are then joined in tandem to form single chimeric sequences A and B. This means that for each 55 target used in the molecular test, standard A would act as positive control, while standard B would act as 56 negative control, or vice-versa. Furthermore, the equally partitioning of target sites between standards A 57 and B enables cross-validation of positive and negative controls, increasing the confidence in test results.

Among the benefits of this design, a chimera allows concurrent testing of disparate target regions of a 59 single pathogen or even different organisms and splitting targets between standards A and B enables 60 control cross-validation, facilitating the distinction of control failure from test failure or success.

The recent emergence of the SARS-CoV-2 pandemic has required widespread diagnostic testing for active 62 virus infections, including genome sequencing, but predominantly using real-time reverse-transcriptase 63 polymerase chain reaction (real-time RT-PCR)-based assays [2] [3] [4] . The World Health Organisation (WHO) 64 published seven diagnostic testing protocols for detection of SARS-CoV-2 that have been rapidly adopted 65 worldwide, with over 20 million molecular tests performed globally by mid-2020 [5, 6] 

The different targeted regions for standards A and B were shuffled and joined together to form chimeric sequences.

The paired design of chimeric A/B standards, where a target in A is absent in B (and vice versa), enables the synthetic 86 RNA transcripts to simultaneously act as positive and negative controls for the real-time RT-PCR primer/probe sets. 

DNA synthesis enables rapid and flexible assembly of reference standards, including sequences not 92 present in natural organisms. This allows sequences from different genome regions to be aggregated to 93 address specific requirements in a diagnostic assay. To demonstrate this approach, we designed synthetic 94 reference sequences that encompass the primer binding sites of all WHO-published real-time RT-PCR 95 tests.

We first retrieved the SARS-CoV-2 genome sequence (isolate Wuhan-Hu-1, NC_045512.2), as well as the 97 primer sequences published by the World Health Organisation (WHO) for China, Hong Kong, Thailand,

United States (CDC), Germany and France (Fig. 1a) . Each available real-time RT-PCR test typically 99 comprises 2-3 primer pairs that target different regions of the SARS-CoV-2 genome (see Supplementary 100 Table 1 ). We then aligned the primer pairs to the SARS-CoV-2 genome and identified the coordinates of 101 the amplicons, which were then retrieved along with an additional 30 nucleotides (nt) on either flanking 102 side (Fig. 1a) .

We next organised these sequences across two different controls (termed chimeric A/B standards). We 104 partitioned the different targeted regions used by each country into two independent groups and then 105 assembled the regions in tandem (Fig. 1b) . A fragment of the human RNase P gene (RP), which is used as 106 a positive human control, was also added to the chimeric standard B.

An additional unique control sequence (UCS) was also included at the 5' region of each standard to enable 108 the unique detection of the standards if required. Each standard sequence was then preceded by a T7 109 promoter to enable in vitro transcription, and followed by a poly-A tract (30nt length) and a restriction 110 site (EcoR1) to enable vector linearization.

The two distinct chimeric A/B standard sequences were then synthesised and cloned into pMK vector 112 backbones (see Methods). We then linearized the plasmids and performed in vitro transcription to 113 produce the synthetic RNA standards (Fig. S1 ). The resulting RNA transcript products were then purified 

Chimeric standards compared to RNA from confirmed COVID-19 RNA patient samples.

We next validated the chimeric A/B standards by comparison to alternative reference controls. We first 127 performed real-time RT-PCR test using the established CDC primers and protocol [10] . Specifically, this 128 employs CDC primers (IDT) N1, N2 and N3 targeting the N gene from SARS-CoV-2 and the human RNase P 129 gene. Chimeric standard A includes regions of the N1 and N3 targets, while chimeric standard B includes 130 regions of the N2 and RP targets. This allows for cross-validation between the chimeric A/B standards, 131 since the standards alternatively act as positive and negative controls to each primer/probe set in the 132 real-time RT-PCR test.

The real-time RT-PCR was initially performed on the chimeric controls alone. We prepared 10-fold 134 dilutions for each control, starting at 3.96 x 10 8 copies/μl for A and 4.22 x 10 8 copies/μl for B (see 135 Methods). As anticipated, in the reactions containing standard A, N1 and N3 primers returned positive 136 results, while N2 and RP were undetected (Fig. 2a) . In contrast, in reactions containing standard B, N2 and 137 RP primers returned positive results, while N1 and N3 were undetected (Fig. 2a) (Figs. 2c,d) . However,

for the RP target gene, which is a positive control for human samples, the Ct value for patient cDNA was 

Observed Ct values for targets N1, N2, N3 and RP at different dilutions of standards A and B (10 0 , 10 -2 , 10 -4 and 10 -173 6 ), across three technical replicates.

To determine the limit of detection (LoD) for the chimeric A/B standards, we performed 100-fold serial 176 dilutions with three technical replicates. As a baseline, we spiked the standards A (3,964 copies/μl) and B

(4,221 copies/μl) into separate background samples consisting of the human universal RNA (100 ng). We 178 then performed real-time RT-PCR using the CDC protocol (see Methods). As a result, we detected both 179 standards A and B until 10 -4 dilution, which corresponds to approximate LoD of 0.39 and 0.42 copies/μl, 180 respectively (Fig. 3a) . As a positive control, the RP primer targeting the human RNase P gene were 181 successfully detected in all tested dilutions, for both standards A and B (Fig. 3b) . Interestingly, N1 primers 182 appear to be more efficient than N3 primers in estimating the LoD for standard A, since the Ct values for 183 N1 (10 0 =21.75 ± 0.09, 10 -2 = 28.26 ± 0.04 and 10 -4 = 34.33 ± 0.96) are significantly lower than N3 (10 0 =25.04 184 ± 0.06, 10 -2 = 31.33 ± 0.04 and 10 -4 = 37.16), in every dilution, across replicates (Fig. 3b) 

The advent of routine DNA synthesis has enabled rapid provision of synthetic reference standards that 189 can be used to validate the accuracy of diagnostic tests. The synthesis of DNA provides a flexible platform 190 to manufacture different reference standards, including non-natural designs. In this case, a single 191 standard can be designed to contain distant genomic regions or even sequences from different organisms.

This allows multiple sequences of interest to be included and organised within a single chimeric standard 193 according to the specific requirements of a diagnostic assay. In this study, we use this approach to 

The A/B standards were synthesized by a commercial vendor (ThermoFisher -GeneArt) and cloned into 267 pMK vectors. The plasmids containing the standards were each resuspended in 50 μl nuclease free water 268 and transformed in E. coli as per manufacturer's protocol (α-Select Competent Cells, Bioline, Australia).

The transformed cells were grown overnight (37°) in LB agar plate containing Kanamycin (100 μg/ml), 

Quantitative real time PCR

Twenty µl reactions were prepared containing 5 µl of input RNA (patient samples, A/B standards or IDT 291 controls), 5 ul of TaqPath TM 1-Step RT-qPCR Master Mix, 1.5 µl of the combined CDC primers/probe set 292 and 8.5 µl of Nuclease-free water. Thermo cycling was performed at 25° for 2 min to allow UNG 293 incubation, followed by 15 min at 50° for reverse transcription, then 2 min at 95° for enzyme activation 294 and finally 45 amplification cycles at 95° for 3 seconds and 55° for 30 seconds. The experiment was 295 performed on QuantStudio 7 Flex real-time PCR systems (Thermo Fisher).

We performed two different serial dilution experiments with the chimeric A/B standards. The first was a 300 10-fold serial dilution of standards A and B alone, to test their performance in the real-time RT-PCR assay.

The baseline concentration was 3.96 x 10 8 copies/μl for standard A and 4.22 x 10 8 copies/μl for standard 302 B. We diluted each standard until 10 -5 . The second serial dilution was to estimate the limit of detection 303 (LoD) for the chimeric A/B standards. For this experiment, standards A and B were individually spiked into 304 universal human RNA samples. The baseline concentration was 3,964 copies/μl for standard A and 4,221 305 copies/μl for standard B and they were each added into 100 ng of universal human RNA. We made 100-306 fold dilutions for the A/B standards until 10 -6 , and the experiment was performed in three technical 

Reference standards for next-generation 315 sequencing

A novel coronavirus from patients with pneumonia in China

A pneumonia outbreak associated with a new coronavirus of probable bat 319 origin. nature

Tracking the COVID-19 pandemic in Australia using genomics. 321 medRxiv, 2020. 322 5. WHO. Molecular assays to diagnose COVID-19: Summary table of available protocols

Rapid establishment of laboratory diagnostics for the novel coronavirus 328 SARS-CoV-2 in Bavaria

Detection of 2019 novel coronavirus (2019-nCoV) by real-time RT-331 PCR. Eurosurveillance

Molecular diagnosis of a novel coronavirus (2019-nCoV) causing an 333 outbreak of pneumonia

An emergent clade of SARS-CoV-2 linked to returned travellers from 337 Iran. Virus Evolution