Highly Sensitive Lineage Discrimination of SARS-CoV-2 Variants through Allele-Specific Probe PCR

Abstract

Tools to detect SARS-CoV-2 variants of concern and track the ongoing evolution of the virus are necessary to support public health efforts and the design and evaluation of novel COVID-19 therapeutics and vaccines. Although next-generation sequencing (NGS) has been adopted as the gold standard method for discriminating SARS-CoV-2 lineages, alternative methods may be required when processing samples with low viral loads or low RNA quality. To this aim, an allele-specific probe PCR (ASP-PCR) targeting lineage-specific single nucleotide polymorphisms (SNPs) was developed and used to screen 1,082 samples from two clinical trials in the United Kingdom and Brazil. Probit regression models were developed to compare ASP-PCR performance against 1,771 NGS results for the same cohorts. Individual SNPs were shown to readily identify specific variants of concern. ASP-PCR was shown to discriminate SARS-CoV-2 lineages with a higher likelihood than NGS over a wide range of viral loads. The comparative advantage for ASP-PCR over NGS was most pronounced in samples with cycle threshold (C_T) values between 26 and 30 and in samples that showed evidence of degradation. Results for samples screened by ASP-PCR and NGS showed 99% concordant results. ASP-PCR is well suited to augment but not replace NGS. The method can differentiate SARS-CoV-2 lineages with high accuracy and would be best deployed to screen samples with lower viral loads or that may suffer from degradation. Future work should investigate further destabilization from primer-target base mismatch through altered oligonucleotide chemistry or chemical additives.

Keywords: SARS-CoV-2; allele-specific probe PCR; diagnostics; next-generation sequencing; variant identification; variants of concern.

FIG 1

Performance of ASP-PCR and NGS in REMAP-CAP trial. (A to C) Individual method performance on REMAP-CAP samples. Integers indicate total samples tested in 1-log data bins. (D) Probit regression of likelihood of lineage designation success for ASP-PCR, NGS-SNP, and NGS-Pangolin with 95% confidence intervals derived from REMAP-CAP samples.

FIG 2

Performance of ASP-PCR and NGS in COV003 trial. (A to C) Individual method performance on COV003 samples. Numbers above indicate total samples in 1-log data bins. (A) Probit regression of likelihood of lineage designation success for ASP-PCR, NGS-SNP, and NGS-Pangolin with 95% confidence intervals derived from COV003 samples.

FIG 3

Impact of degraded RNA in COV003 samples on method performance. (A) Genome coverage of COV003 samples plotted versus sample viral load. Samples with ≤20,000 bases with ≥2 reads and viral loads of >10⁶ IU/mL were defined as degraded (red box). (B to D) Individual method performance on COV003 degraded and nondegraded samples. Numbers above indicate total samples in 1-log data bins.

FIG 4

Effect of molecular modification on S:D1118H/Alpha ASP-PCR performance. Percentage of total samples successfully typed using approach described in Materials and Methods. P values from Fisher’s exact test. Lineage designations (e.g., wild type [WT] or Alpha/B.1.1.7) were concordant for all samples for all methods.