key: cord-0266512-4vl8zpev
authors: Spiess, K.; Gunalan, V.; Marving, E.; Nielsen, S. H.; Joergensen, M. G. P.; Fomsgaard, A. S.; Nielsen, L.; Alfaro-Nunez, A.; Karst, S. M.; Mortensen, S.; The Danish COVID-19 Genome Consortium,; Rasmussen, M.; Lassauniere, R.; Rosenstierne, M. W.; Polacek, C.; Fonager, J.; Cohen, A. S.; Nielsen, C.; Fomsgaard, A.
title: Rapid surveillance platforms for key SARS-CoV-2 mutations in Denmark
date: 2021-10-26
journal: nan
DOI: 10.1101/2021.10.25.21265484
sha: 0868350b9d4536df0b561088d59ee9d48597fbb4
doc_id: 266512
cord_uid: 4vl8zpev

Multiple mutations in SARS-CoV-2 variants of concern (VOCs) may increase, transmission, disease severity, immune evasion and facilitate zoonotic or anthoprozoonotic infections. Four such mutations, {Delta}H69/V70, L452R, E484K and N501Y, occur in the SARS-CoV-2 spike glycoprotein in combinations that allow detection of the most important VOCs. Here we present two flexible RT-qPCR platforms for small- and large-scale screening to detect these mutations, and schemes for adapting the platforms for future mutations. The large-scale RT-qPCR platform, was validated by pair-wise matching of RT-qPCR results with WGS consensus genomes, showing high specificity and sensitivity. Detection of mutations using this platform served as an important interventive measure for the Danish public health system to delay the emergence of VOCs and to gain time for vaccine administration. Both platforms are valuable tools for WGS-lean laboratories, as well for complementing WGS to support rapid control of local transmission chains worldwide.

The global SARS-CoV-2 pandemic, which raised with the identification of this novel 50 coronavirus in late 2019, has seen the emergence of several variants, each with a 51 distinct set of mutations 1 . Early detection of new SARS-CoV-2 mutations and 52 associated measures to decrease the risk of spread are important to control local 53 outbreaks of SARS-CoV-2 variants, especially those which have been designated 54

Variants of Concern (VOCs) 2,3 . The latter are defined by increased transmissibility, 55

severity of infections and resistance to immunity [4] [5] [6] [7] [8] * former variant of concern, now classified as de-escalated variant/ adapted from 5 The N501Y mutation occurs in the receptor-binding interface and confers a 69 substantial increase in the binding affinity of the S for the human angiotensin-70

converting enzyme 2 (hACE2) protein 17 . HACE2 interaction with the S is essential for 71

virus entry and infection of the cells 18 . The E484K mutation has been identified as an 72

immunodominant spike protein residue, facilitating escape from several monoclonal 73 antibodies, as well as antibodies in convalescent plasma [19] [20] [21] . Altered immune 74 recognition has also been described for the L452R mutation 21-23 . The key mutation 75 ΔH69/V70, a two amino acid deletion, has appeared in multiple SARS-CoV-2 76 variants at different geographical locations across Europe. In Denmark, ΔH69/V70 77 was detected in local outbreaks in mink farms in Northern Jutland 24,25 . The spread of 78 this deletion in combination with additional mutations (notably Y453F) resulted in the 79 SARS-CoV-2 mink variants B.1.1.298, which transmitted both ways between humans 80 and mink; also giving rise to the early "cluster 5" variant 24,26. .

The identification of these variants and the mutations that form their signature are 83 largely dependent on Whole Genome Sequencing (WGS) of SARS-CoV-2 from 84 infected individuals. In addition, WGS of SARS-CoV-2 also elucidates sets of novel 85 mutations potentially linked to changes in viral properties or associated with vaccine 86

breakthrough. However, the utility of WGS in a pandemic such as this also carries 87

with it a significant cost in the form of reagents, equipment as well as turnaround time 88 nucleotide sequence was detected independent of the concentration of the SARS-139

CoV-2 sample (amount of SARS-CoV-2 RNA included per sample into the PCR) 140 (Fig. 2F) for the ΔH69/V70 deletion (Fig. 2I) , which was confirmed by WGS (Fig. 2H , Suppl. Fig 1 D) . As a first step, the ΔH69/V70 RT-qPCR ran 156

together with the diagnostic SARS-CoV-2 E-Sarbeco PCR (E-gene) 29 . The sensitivity 157

of the ΔH69/V70 RT-qPCR was found not to be reduced when multiplexed with the 158 E-Sarbeco RT-qPCR (Suppl. Fig. 1E ). In conclusion, the ΔH69/V70 RT-qPCR was 159 determined to be sensitive and specific for the detection of the ΔH69/70 as well as 160

insensitive to multiplexing. To detect further key-mutations present in SARS-VOCs 161

(Alpha/Beta/Gamma/Delta) and other variants of interest, we developed primers and 162

probes to detect the L452R, E484K and N501Y mutations. Compared to the 163 ΔH69/V70 deletion where the probe targets a stretch of a deletion of six nucleotides, 164

the probes for the three key mutations listed above differ only by one nucleotide 165 substitution within the S. Therefore, we increased their binding affinity to the 166 mutations or the WT sequence by modifying the probes as black whole quencher 167 plus-(BHQplus), locked nucleic acid-(LNA) or minor grove binding (MGB) 168 conjugated probes. For the different RT-qPCRs, we tested for each mutations all 169 primer and probe combinations, with all three probe modifications. For the N501Y 170 mutation e.g., the MGB-conjugated probes for the N501Y mutation in the RT-qPCR 171

were observed to be superior to locked nucleic acid (LNA) -conjugated probes, where 172 a specific signal was detected for either the mutation or WT sequence. In contrast, the 173 LNA probes in the N501Y RT-qPCR detected the right mutations present in the 174

variants, but additional allelic discrimination analysis was needed to discriminate 175 between the intensity of the signal for the mutation or the WT probe at a Ct of 45 176 (Suppl. Fig. 2A-D) .

For the L452R mutation, BHQ plus conjugated probes were found to be absolutely 178 specific compared to the LNA-and MGB conjugated probes ( Fig. 3A-B) . The limit of 179 detection for L452R was determined by a dilution series of a patient sample with 180 known sequence information for the delta variant (B.1.617.2) and tested in parallel in 181 the L452R RT-qPCR and the E-Sarbeco RT-qPCR (Fig. 3C) . The L452R RT-qPCR 182 was about 2-fold less sensitive than the E-Sarbeco RT-qPCR (Fig. 3C) .

The best results for the E484K mutation were gained using MGB -conjugated probes 184

that were refined to generate a signal specific to mutation or WT nucleotide, 185 respectively ( Fig. 3D-E, G-H) . The limit of detection for the E484K RT PCR was 186 found to be 52 and 5 copies/µl respectively, performing a dilution series with the 187 TWIST control Beta B.1.351 and Gamma P.1 (Fig. 3F , I, Suppl. Tab.1). 188 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in As a signal detected was specific either for the key mutations or the WT sequence, it  190  was possible to only include the probes detecting the key mutations (L452R, E484K  191 and N501Y) or the ΔH69/V70 into the multiplexed RT-qPCRs v.1 and v.2 ( Fig. 3J The same primer and probes designed for the four key mutations (ΔH69/V70, L452R, 234

E484K and N501Y) included into the multiplexed RT-qPCR for small-scale screening 235

were further validated for large-scale screening, implemented to support the national 236

surveillance program in Denmark in addition to WGS, supporting the public health 237 system to delay the emergence of VOC. Large-scale screening consisted of RT-238 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in test of the RT-qPCR strategy due to the presence and absence of these key mutations 260 amongst these multiple variants ( Fig. 1A-B) . A daily range of 150 to 671 samples 261

were analysed by multiplex RT-qPCR during this period and results were 262 characterised as either positive (POS) or negative (NEG) for a given key mutation 263 (Fig. 4C) . It was observed that there was a small number of inconclusive results 264 amongst the E484K RT-qPCRs, as well as a more noticeable number of inconclusive 265 results in the N501Y RT-qPCR which could be attributed to probe manufacturing 266 issues beyond our controlreplacement of the probes resulted in a significant 267 reduction in the number of inconclusive results from this reaction ( Fig. 4C/ four key mutations for various technical reasons anticipated at this scale (see above). 284

The validation results (Fig. 5) showed good agreement between amino acids 285 translated from WGS and RT-qPCR results for E484K, N501Y and L452R ( Fig. 5A-286 C). The determination of concordance proved less straightforward for ∆H69/V70 due 287

to the alignment of reads around the deletion prior to consensus generation, resulting 288 in a significant discordant fraction between the deletion and negative RT-qPCR 289 results (Fig. 5D) . It was also observed amongst the consensus genomes used in this 290

validation that amino acid 452 in the spike protein was more mutable than the other 291 positions which form this set of key mutations, with L, R, M and Q observed at this 292 position depending on the lineage (Q484 was not observed in genomes during the 293 selected period but has been recorded in global surveillance data).

In order to meaningfully compare and describe the relative performance of the RT-296 qPCR strategy from the results of the large-scale screen, as well as to determine the 297 true-positive or true-negative rate of these combinations of primers and probes, the 298 specificity and sensitivity of each primer/probe combination was determined using 299

established methods used to characterise diagnostic testing 32 . Using the validation of 300

the RT-qPCR with WGS as a reference standard, specificities and sensitivities were 301 calculated for all primers and probes for the four key mutations, and it was observed 302 that all four RT-qPCRs were highly specific (>99.9%), and three out of four assays 303

were highly sensitive (>99.9%) ( Fig. 5 A-D) . The sensitivity of the probe for 304 detection of ∆H69/V70 was observed to be reduced (79.28%) due to a significant 305 number of deletions in WGS, which were assayed to be negative by RT-qPCR; 306 however, given the challenge of read alignments around genomic regions containing 307

insertions or deletions, this was postulated to be largely due to the determination of 308 the deletion in WGS consensus genomes. In addition to the specificity and sensitivity 309

of the primer/probe combinations, the Positive Predictive Value (PPV) and Negative 310

Predictive Value (NPV) of these combinations was also determined, which indicates 311 the ability of a diagnostic assay or test to accurately detect a condition or in this case, 312 mutation 32 . The determination of PPV and NPV takes into account the specificity and 313 sensitivity of the primer/probe combinations as well as the prevalence of the four key 314 mutations amongst the sequenced SARS-CoV-2 genomes during the study period: 315

It was determined that all four primer/probe combinations had a PPV of at least 97 %, 316

and a NPV of 99.9 % for three out of four of these, with the ∆H69/V70 assay having a 317 NPV of 75.43 % (Fig. 5 A-D) . From the results of the large-scale screening, it can be 318 seen therefore that the specificity and sensitivity as well as the PPV and NPV all point 319 towards the viability of this RT-qPCR strategy in a large-scale diagnostic setting. 

In summary, we developed a RT-qPCR system for large-scale screening of four key 327 mutations in parallel that is highly specific and sensitive, validated by a comparison of 328 the qPCR and WGS data from 9572 samples that were tested in parallel. 329 330 DISCUSSION 331 332 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted October 26, 2021. ;

RT-qPCRs platforms for small and large-scale screening can support the detection of 333 mutations of concern present in SARS-CoV-2 variants. This is of special interest for 334 countries lacking an infrastructure for large-scale WGS sequencing, the golden 335 standard for SARS-CoV-2 variant surveillance. Here, a detection system is needed 336 that is fast, robust and flexible and, which enables the detection of known diagnostic 337 mutations almost in real-time after sample collection, as we showed in this study. 338

Here, we describe validated and advanced RT-qPCR platforms for small and large-339 scale screening that can simultaneously detect mutations of concern within the S of 340 SARS-CoV-2, with a fast turnaround time for large-scale screening of 12-24h to 341

report to the public health system. In comparison to commercially available systems 342

to detect mutations of concern, the RT-qPCR platforms can be established fast and 343 new mutations can be implemented; an important advantage to follow the course of a 344

pandemic. It is a transparent system, where troubleshooting is possible without 345 depending on the knowledge from a company and it can be adjusted to the existing 346

infrastructure of the laboratory for large-scale screening and data evaluation. 347

Moreover, the RT-qPCR platforms are at low cost of about 10 DKK (2 USD) per 348 reaction, and can therefore be establish in countries without the resources for WGS in 349

large-scale 350

RT-qPCR is a fast, standard method for SARS-CoV-2 detection and has been 351 established at the start of the pandemic in January 2020 29 . The standardised protocol 352

for SARS-CoV-2 RT-qPCRs makes it easy to implement into diagnostic laboratories 353 worldwide, where the equipment needed is commonly present. This could be of 354 advantage compared to new methods such as RT-LAMP or CRISPR, which have 355 been described for SARS-CoV-2 detection, delivering faster test results and can be 356

applied without extensive laboratory equipment as RT-qPCRs 33-35 .

Currently there are only limited studies on RT-LAMP for commercial point of care 3 . 358

Moreover, CRISPR is still in its infancy 36 and has been shown to be less sensitive 359 compared to RT-qPCR 3 . As most diagnostic facilities worldwide do not possess 360

access and knowledge to establish these technologies, opposing to RT-qPCR that is a 361 universal standard method, RT-qPCRs are still the method of choice for most 362 diagnostic laboratories. Based on recent advances in modifications of conjugated-363

probes, RT-qPCRs can be designed to detect mutations within the SARS-CoV-2 364 genome consisting out of a single nucleotide polymorphism (SNP). 365

For small-scale screening the multiplexed RT-qPCR was developed using a Luna 366

Probe One-Step RT-qPCR Mix, which offers the possibility to increase the input 367 template concentration, for amplification targets with a low RNA concentration, as it 368

is four times more concentrated. However, this was not an advantage when 369 establishing the multiplexed RT-qPCRs v.1 and v.2, as SARS-CoV-2 RNA 370

concentrations vary among patient samples and can be too high from start leading to 371 artificial signals. In contrast, adjusting the primer and probe concentrations for each 372 mutation resulted in a highly specific and sensitive detection of the corresponding 373 mutation present in the different SARS-CoV-2 variants ( Fig. 3J-O, Suppl. Tab. 4) . 374 Moreover, by reducing the number of probes in the multiplexed RT-qPCR we could 375 maintain a sensitive system for diagnostic use, by including one probes for each 376 mutation. This was possible by designing and testing combinations of primer and 377

MGB-, LNA-or BHQplus-conjugated probes that yield a specific signal for the 378 mutation and WT sequences, respectively. The best performance for each 379 primer/probe pairs is empirical and should be tested for all possible probe 380 modifications (LNA, MGB and BHQplus), as the result is dependent on the 381 nucleotide sequence of the mutation or WT sequence. So far, only mutations of 382 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted October 26, 2021. ;

concern within the S were included into the multiplexed RT-qPCR platform for small-383 scale screening but running the ΔH69/70 RT-qPCR as multiplexed PCR together with 384

the clinical E-sarbeco RT-qPCR did not reduce the sensitivity of the PCR (Fig. 2E) . 385 As up to five targets can be included into the multiplex RT-qPCR using the Luna 386

Probe One-Step RT-qPCR Mix, additional targets located in other loci of the SARS-387

CoV2 genome than S could be included. We did not test the maximum number of 388 targets that could be included into RT-qPCR platform, as this was out of the scope of 389 this study, but this could be of interest for future studies. 390

Large-scale RT-qPCR screening of mutations present in VOCs required that a number 391 of technical and analytical considerations can be fulfilled: 1) the RT-qPCR must be 392

highly specific and sensitive to minimise or avoid false positives, 2) it should be of 393 sufficient robustness to allow for massive scalability required in a pandemic, 3) it 394

must not interfere with the diagnostic PCR to detect SARS-CoV-2 to reduce the risk 395

of potential PCR contamination 4) it requires liquid handlers in order to be viable 396 from a practicable standpoint and 5) an advanced, automated evaluation system is 397

needed to detect the erroneous results. Here we describe a large-scale RT-qPCR 398

platform that meets the technical and analytical considerations outlined above. The 399 current design is based on sample preparation in a 96-well format and subsequent RT-400 qPCR in 384-format. This allows each sample to be analysed by four separate sets of 401 primers and probes, which enables the analysis of four mutations for up to 92 samples 402

and four controls (one negative and three positive) in parallel in a single run. The 403

system is flexible as the combination of target mutations can be adjusted over time in 404

accordance with current needs. The handling of data calls for automated data 405

processing and variant calling which is due the large amount of data in each run and 406

the complex calling algorithms. Inconclusive results can pose a challenge with regards 407

to variant calling. When one or more of the mutations are inconclusive, it is not 408 always possible to make an unequivocal variant call. In our set-up we have opted to 409 report the detected mutations. In these cases, prominent mutations with putative 410 biological functions in various VOCs were reported, rather than variants of concern 411 and interest. 412

From the large scale-screen it was determined that the RT-qPCR platform described 413

in this study is generally of very high specificity and sensitivity and performs well in 414 terms of its PPV and NPV, indicating its utility in such large-scale diagnostic screens. 415

The period for the large-scale screening and validation was specifically chosen to 416

interrogate the robustness of this system in a pandemic transition period with ongoing 417 lineage replacement; such a period involves the waning of certain variants such as the 418

Alpha (B. not a rare occurrence, and a certain degree of N-counts is therefore considered 431 permissible (typically less than 5-10 % of the consensus genome). Tracts of Ns 432 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted October 26, 2021. ; around this region were observed around the deletion and this was largely responsible 433

for the challenges of identifying a deletion from WGS consensus genomes. However, 434

this was not the case for single SNPs leading to non-synonymous substitutions as seen 435

with N501Y, E484K and L452R. Interesting insights into the specificity and the 436 sensitivity of the RT-qPCR system were also observed in the results around the 437 L452R mutation, given that position L452 in the spike protein exhibited more than a 438 single amino acid change during the pandemic and indeed the timeframe of the large-439

scale screening performed. The validation showed that all samples with L452Q in the 440 spike protein recorded a positive result from the RT-qPCR whereas L452M 441 exclusively recorded negative RT-qPCR results. Given that the codon observed from 442 WGS encoding for Q was cag and the corresponding codon encoding for L at the 443 same position was cgg, this could be considered unsurprising, also given that Q452 444

was not an anticipated mutation and therefore was not considered in the design of the 445 probes. Given that the codon atg, which is more distant from cgg and which encodes 446

for M at this position, was not detected by the L452R-specific probe, this alludes to 447 the specificity and sensitivity of the RT-qPCR probe at position 452. 448 449

One of the major arms of pandemic control seen in this pandemic revolves around the 450 screening and isolation of SARS-CoV-2 infected individuals in order to limit 451 community spread of infections. The screening and isolation of individuals is 452 therefore time sensitive and requires a rapid turnaround, especially where the control 453

of variants or mutations of concern are a priority. While WGS of positive samples 454

affords the accurate identification of these variants or mutations in order to enable 455 their tracking and therefore control, this entails a longer turnaround time and greater 456 cost in terms of reagents, equipment and expertise. The use of RT-qPCR systems such 457

as the one described in this study allows for rapid identification of mutations of 458 concern, which in turn enables near-real-time tracking of these and correspondingly, 459 rapid decision-making around testing, contact tracing and isolation. This enabled the 460 rapid reaction of the public health system in Denmark to the detection of VOCs, with 461 the added benefit of gaining time to implement its vaccination schedule; being in line 462

with modelling showing that minimising testing delay, had the largest impact on 463

reducing onward transmissions 37 . The flexibility of this system also allows for 464 multiplexing to detect multiple mutations and the incorporation of new primers and 465

probes in response to the dynamics of the SARS-CoV-2 pandemic. In addition, the 466 specificity and sensitivity of this system show that it is robust and therefore suited to 467 diagnostic requirements in a pandemic. Taken together, these characteristics make this 468 RT-qPCR system an ideal candidate for laboratories looking to detect mutations of 469 concern in the SARS-CoV-2 pandemic. 

All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in were aliquoted and frozen at -80°C. Subsequent passages to expand virus stocks were 502 performed in 75 cm 2 flasks seeded with 1.5x10 6 Vero E6 the day before. 25 µL of 503 primary isolate supernatant was used as inoculum in the presence of 2 mL infection 504 media. After 1h at 37°C/5% CO2 incubation, flasks were supplemented with 10 mL of 505

propagation media (without Amphotericin) and incubated until CPE was obtained. 506

Supernatants were then clarified by centrifugation for 5 min at 300 x g and stored as 507

single use aliquots at -80°C. 508 509

For determining specificity and sensitivity of the SARS-CoV-2 Variant PCR assays, 511 the following materials were used: 512

Diagnostic samples positive for the common respiratory pathogens Human input and 100 µL elution volume). 547

For positive controls, 120 µL of supernatant from SARS-CoV-2 infected cells were 548 mixed with 120 µL of MagNA Pure lysis buffer (Roche) and extracted as small-scale 549 SARS-CoV-2 patient samples. Positive control RNA was stored at -80°C until use. 550

For large-scale SARS-CoV-2 patient sample screening, RNA was extracted using a 551

Beckman Coulter Biomek i7 robot using the Beckman Coulter RNAdvance Whole 552 blood kit (200 µL input and 50 µL elution volume).

Primer and probe design 555 SARS-CoV-2 variant sequences were retrieved from positive samples identified 556 through the national surveillance program in Denmark. Sequences were aligned and 557

primer and probes were designed using Geneious Prime 2021.0. 558

Two probes were designed for each key mutation: one detecting the wildtype (WT) 559

nucleotide sequence, and one detecting the mutation. The probe design was refined to 560 detect the key mutations (L452R, E484K, N501Y, Δ69/V70 deletion) with only one 561 probe in the multiplex RT-qPCRs. To ensure stable allelic discrimination analysis, 562

probes detecting the mutations with only one nucleotide exchange were either MGB, 563

LNA or BHQplus modified, which increases the melting temperature (Tm). The 564 calculation of MGB probe Tm was adapted from 38 . 565 566

The primers and probes listed in Tab. 2 were synthesized by Biosearch Technologies, 567

Denmark, except for the MGB-probes that were synthesized by Eurogentec, Belgium, 568 and the Zen-probe, that was synthesized by Integrated DNA Technologies, Belgium. 569

All oligos were HPLC-purified. perpetuity.

preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in 

(1) While it is more common to use a BHQ1 quencher together with HEX, this system works well 577 with a BHQ2 quencher.

(2) Volumes of oligos added to mastermix are valid for both 96 and 384-well formats.

Mastermix set-up 581

The primers and probes were combined in different master-mixes. 582

In master-mix 1-4 (large-scale screening): the mutations were detected using both the 583 mutant probe and the wildtype probe for allelic discrimination analysis. 584

In master-mix 5 and 6 (small-scale screening), only probes targeting the mutations 585

were used, and therefore no allelic discrimination analysis was needed. 586 587 96-well format PCR conditions used in development phase and small-scale testing 588

All PCR assays were developed on a Bio-Rad CFX 96 PCR real-time PCR system. 589

Master-mix 1 -4 contained 12.5 µL Luna ® Universal Probe One-step RT-qPCR Kit 590 reaction buffer (NEB), 1.25 µL Luna ® WarmStart RT Enzyme mix, primers and 591 probes (100 µM, volumes in table 1), DNAse/RNAse free water and 5 µL template to 592 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted October 26, 2021. ; a total volume of 25 µL. Cycling conditions: Reverse transcription at 55 0 C for 10 593 min., initial denaturation at 95 0 C for 3 min., followed by 45 cycles of denaturation 594 and annealing/extension at 95 0 C for 15 sec. and 58 0 C at 30 sec. respectively. 595

Master-mix 5 -6 contained 5 µL Luna Probe One-Step RT-qPCR 4X Mix with UDG 596

(New England Biolabs Inc (NEB)), primers and probes (100 µM, volumes in Tab. 1), 597

DNAse/RNAse free water and 5µL template to a total volume of 25 µL. Cycling 598 conditions: Initial step at 25 0 C for 30 sec, reverse transcription at 55 0 C for 10 min., 599 initial denaturation at 95 0 C for 1 min., followed by 45 cycles of denaturation and 600

annealing/extension at 95 0 C for 10 sec and 58 0 C at 60 sec respectively. 601 602

Data analysis for the multiplexed RT-qPCRs used in small-scale testing 603

The multiplexed RT-qPCRs contain probes only targeting the mutations, ΔH69/V70, 604 501Y, 484K for master-mix 5, and 501Y, 484K, 452R for master-mix 6. Cut-off 605 values were used in the multiplexed RT-qPCRs to secure the detection of only the 606 mutation and not the WT sequence as there was no WT probe in the mix. A sample 607

was considered positive with these criteria: Ct <38 and RFU (Relative Fluorescence 608

Units) > 500 at Ct = 45. 609 610

384-well format PCR conditions for large-scale testing 611

In large scale testing, the assays run on a Bio-Rad CFX 384 PCR real-time PCR 612

system. The master-mix contained 7.5 µL Luna ® Universal Probe One-step RT-qPCR 613

Kit reaction buffer (New England Biolabs Inc), 0.75 µL Luna ® WarmStart RT 614

Enzyme mix, primers and probes (100 µM, volumes in table 1), DNAse/RNAse free 615 water and 5µL template to a total volume of 15 µL. Cycling conditions were the same 616 as for the 96-well format. Each patient sample was analysed in four PCR wells, in 617 four parallel reactions, using master-mix 1 or 2 for detecting ΔH69/V70 and N501Y, 618 master-mix 3 for detecting E484K, master-mix 4 for detecting L452R and in the final 619

well the E-Sarbeco assay was used for detection of SARS-CoV-2 wildtype (E-gene). 620

The 4 master-mixes were placed in a quadratic pattern, thus allowing easy transfer 621 from a 96-well plate to a 384-well plate (e.g. A1 in a template plate was pipetted to 622 A1, B1, A2 and B2 of the master mix plate). Master-mix 5 and 6 were not tested in 623 the 384-well format.

Data analysis using allelic discrimination analysis in large-scale testing 626 PCR curves were evaluated in the Bio-RAD CFX software and Ct values and end 627 RFU were exported in csv files. The files were imported into the laboratory database 628

where all data analysis was performed. For the ΔH69/V70 deletion, detection was 629 based on Ct values (deletion detected is Ct = 12-38). For the the mutations N501Y, 630 E484K and L452R, detection was based on allelic discrimination where the end RFU 631

values were utilized to determine the presence of a mutation (see Suppl. Tab. 1). A 632 sample was considered positive with these criteria: Ct <38 and RFU > 200 at Ct = 45. 633

The RFU cut-off value was used in the 384-well PCR-format as a quality control step, 634

in case one of the probes in the allelic discrimination pair failed.

Whole genome sequencing 637

Whole genome sequences were generated by The Danish COVID-19 Genome 638

Consortium (DCGC) from PCR-positive samples collected between 6 th June and 11 th 639

July 2021. Samples were selected using Ct cut off values between 30 -38 30 . The bulk 640 of the samples were sequenced using the ARTIC Network tiled PCR scheme V3 via 641 the COVIDseq Assay [Illumina], Artic Network nCoV-2019 sequencing protocol v2 642 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in We used standard curves to determine the SARS-CoV-2 detection threshold for each 676 assay and to calculate the viral load in each sample. We used the SARS-CoV-2 677 variant specific TWIST control with a known concentration (copies/µl) and diluted 678 1:10 in a seven-step dilution series. The median Ct-values and the interquartile ranges 679

were calculated based on biological duplicates with technical duplicates. The A-B) LNA probes detecting the 501Y mutation and N501 WT sequence respectively. 762 C) Allelic discrimination analysis to differentiate between the 510Y mutation and 763

N501WT sequence. D) Multiplexed PCR to detect the ΔH69/V70 mutation and the 764 N501Y mutation. 765

We would like to extend our gratitude to Susanne Lopez Rasmussen and Halenur A. 767

Bayhan for their assistance and technical support. We would like to acknowledge the perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in

The copyright holder for this this version posted October 26, 2021. 

Deep Mutational Scanning of SARS-CoV-2 Receptor Binding 833 Domain Reveals Constraints on Folding and ACE2 Binding

Viral dynamics in asymptomatic patients with COVID-19

Escape from neutralizing antibodies by SARS-CoV-2 spike protein 838 variants

Prospective mapping of viral mutations that escape antibodies 840 used to treat COVID-19

SARS-CoV-2 variants, spike mutations and immune 842 escape

Identification of SARS-CoV-2 spike mutations that attenuate 844 monoclonal and serum antibody neutralization

Comprehensive mapping of mutations in the SARS-CoV-2 847 receptor-binding domain that affect recognition by polyclonal human plasma 848 antibodies

Monitoring of SARS-CoV-2 infection in mustelids

SARS-CoV-2 transmission between mink (neovison 852 vison) and Humans

Antibody Neutralization of SARS-CoV-2 Cluster 5 Variant Emerging in Mink 855 at Danish Farms

High-throughput Mutational Surveillance of the SARS-CoV-2 857

A rapid, cost efficient and simple method to identify 860 current SARS-CoV-2 variants of concern by Sanger sequencing part of the 861 spike protein gene

Detection of 2019 novel coronavirus (2019-nCoV) by real-time 864 RT-PCR

Introduction and transmission of SARS-CoV-2 B.1.1.7 866 in Denmark

Viral infection and transmission in a large, well-traced outbreak 869 caused by the SARS-CoV-2 Delta variant

Pliabilities, and Pitfalls in Research and Practice. Front. Public Heal. 0

RT-LAMP for rapid diagnosis of coronavirus SARS-CoV-875 2

Clinical validation of optimised RT-LAMP for the diagnosis of 877 SARS-CoV-2 infection

CRISPR-Cas12-based detection of SARS-CoV-2

Sherlock and detectr: CRISPR-cas systems 881 as potential rapid diagnostic tools for emerging infectious diseases

Impact of delays on effectiveness of contact tracing 884 strategies for COVID-19: a modelling study

Efficient priming of PCR with short oligonucleotides 887 conjugated to a minor groove binder

Improvements to the ARTIC multiplex PCR method for 889 SARS-CoV-2 genome sequencing using nanopore

MAFFT Multiple Sequence Alignment Software 891 Version 7: Improvements in Performance and Usability

MAFFT: a novel method for rapid multiple 894 sequence alignment based on fast Fourier transform

Jalview Version 2-a multiple sequence alignment editor and analysis 898 workbench

The authors declare no competing interests 785 786 Insights from linking epidemiological and genetic data.