Rapid and accurate identification of SARS-CoV-2 Omicron variants using droplet digital PCR (RT-ddPCR)

Abstract

Background: Some mutations in the receptor binding domain of the SARS-CoV-2 Spike protein are associated with increased transmission or substantial reductions in vaccine efficacy, including in recently described Omicron subvariants. The changing frequencies of these mutations combined with their differing susceptibility to available therapies have posed significant problems for clinicians and public health professionals.

Objective: To develop an assay capable of rapidly and accurately identifying variants including Omicron in clinical specimens to enable case tracking and/or selection of appropriate clinical treatment.

Study design: Using three duplex RT-ddPCR reactions targeting four amino acids, we tested 419 positive clinical specimens from February to December 2021 during a period of rapidly shifting variant prevalences and compared genotyping results to genome sequences for each sample, determining the sensitivity and specificity of the assay for each variant.

Results: Mutation determinations for 99.7% of detected samples agree with NGS data for those samples, and are accurate despite wide variation in RNA concentration and potential confounding factors like transport medium, presence of additional respiratory viruses, and additional mutations in primer and probe sequences. The assay accurately identified the first 15 Omicron variants in our laboratory including the first Omicron in Washington State and discriminated against S-gene dropout Delta specimen.

Conclusion: We describe an accurate, precise, and specific RT-ddPCR assay for variant detection that remains robust despite being designed prior the emergence of Delta and Omicron variants. The assay can quickly identify mutations in current and past SARS-CoV-2 variants, and can be adapted to future mutations.

Keywords: Allelic discrimination; B.1.1.529.

Fig. 1

Probe binding clearly distinguishes mutations in RT-ddPCR. Droplet amplitude plots (n = 1 per allele) illustrate how FAM fluorescence (X axis) and VIC fluorescence (Y axis) are diagnostic of templates matching the FAM probe (first column, blue droplets) or VIC probe (third column, green droplets) compared to templates with mutations in probe sequences (second column, orange droplets) and to droplets that lack template (fourth column, grey droplets) for each reaction (rows). Bar graphs (final column) show the consistency of amplitudes between specimens (average mean amplitude ± average standard deviation of 190–1900 positive droplets each, n = 4 per allele). Note that in the 452/501 reaction, unlike the 417 and 484 reactions, the two probes have separate targets so droplets may show fluorescence from only one or the other target (droplets lower along the axes) or from both targets (droplets higher along the axes).

Fig. 2

Template concentration does not affect assay accuracy. For each reaction (rows), composite droplet amplitude plots showing the highest-concentration sample of each allele (first column) and lowest-concentration dilution with >10 positive droplets of the same samples (second column) illustrate that amplitudes are diagnostic of template sequences despite wide differences in copy number per droplet. Circles or lines separating droplets from different alleles are for reference on the low-concentration plots. Bar graphs (final column) show this is consistent between specimens (mean amplitude ± standard deviation of highest [dark] and lowest [light] concentration of each specimen, n = 4 per allele).

Fig. 3

Additional UWVL-identified mutations in probe sequence have variable effects on assay accuracy. (A) Mutation A->G in Spike_E484 results in decreased droplet amplitude that is still distinguishable from other alleles. (B) Mutation G->A in Spike_K417T results in decreased amplitude that is barely distinguishable from K417. (C-D) K417T sequences from samples collected within the week beginning each listed date in the world as a whole, the USA as a whole, or by UWVL: (C) total K417T sequences; (D) sequences with K417T encoded by ACA instead of ACG codon.

Fig. 4

Omicron specimens are accurately identified with the assay. Droplet amplitude plots (n = 1 per specimen type) illustrate the different assay results for three categories of newly-collected specimen: non-SGTF (first column), the first SGTF identified at UWVL (second column), and all subsequent SGTF (third column) for each reaction (rows). Bar graphs (final column) show average mean amplitude (± average standard deviation) for samples from each category (n = 5, n = 1, and n = 7 respectively). Variant determination (bottom) based on the assay was confirmed in all cases by whole-genome sequencing.