on PCR, use primers (short DNA sequences) to bind and detect specific virus RNA target sequences. These tests remain one of the most accurate and widely used methods for SARS-CoV-2 diagnosis. However, if a mutation occurs in one of the primer target sequences used in these tests, it is possible that the primer may no longer be able to bind to the target, producing a false negative result. In general, most NAATs are designed to have multiple genetic targets [31]. This means that if a mutation does occur in one test target site, the overall test should still work and produce the correct results. Antigen and antibody tests could also be affected by mutations, but for this to happen the mutation would have to cause an alteration to the protein or physical structure of the virus targeted by the test. 3.1 Impact of variants of concern on diagnostics The most well-characterized VOCs, B.1.1.7, B.1.351 and P.1, have multiple mutations, including several in the spike gene (S-gene). If any of these mutations occur at primer binding sites or affect the structure of viral antigen targets that are detected by the antigen tests, they have the potential to impact the accuracy of diagnostic tests. SARS-CoV-2 variants 8 It is now well established that this is the case for the B.1.1.7 variant, due to the presence of a Δ69/70 mutation in the S-gene. This mutation prevents the primers in certain PCR tests with S-gene targets, notably the ThermoFisher TaqPath test, from binding to the S-gene target. This delivers a negative result for this target, widely known as S-gene target failure, or S-gene drop out [32]. However, the ThermoFisher TaqPath test contains three targets, with two targets still working as expected. Due to this built-in redundancy, there has been no impact on the accuracy of the overall test results. As most tests use multiple genetic targets, it is expected that most tests in widespread use should continue to be accurate in the face of viral evolution. There is no evidence to date that the S-gene mutations in the other VOCs alter the performance of PCR tests, although it will be important for users of tests that target the S-gene to monitor this. The mutations in the other genes of the variants do not appear to have had an impact on diagnostic tests [33]. It is important to monitor the impact of mutations in all genomic regions, as most commercially available tests do not use the S-gene as a primary target. For example, the majority of commercially available antigen tests target the N-protein, and many PCR tests also include an Ngene target [34]. The B.1.1.7 variant has been shown to have no effect on five rapid antigen tests in an evaluation by the UK government, despite mutations in the N-gene which could potentially affect antigen structure [35]. The effect of the N-gene mutations observed in other variants on antigen test detection has not been evaluated, but currently no major changes to test performance are anticipated. Whilst there is potential for the mutations in the variants to impact antibody tests, this has not yet been evaluated for any of the variants [36]. 3.2 How the impact on diagnostics has been managed In terms of the impact of S-gene target failure in samples containing the B.1.1.7 variant, in many PCR diagnostic tests the S-gene is not used as a target, so many tests are unaffected by S-gene failure. In tests which do target the S-gene, it is typically only one of several targets, so overall test performance has not been affected. As two working test targets are perceived as sufficient, the ThermoFisher TaqPath test is still used. The overall impact on diagnostic test performance has therefore been minimal; however, the case of S-gene target failure does highlight the need for testing at least two independent targets, as recommended by WHO [37]. Thus, any diagnostic tests that currently have only one additional target may need to be redesigned. The S-gene should be avoided as a target in the future, due to the high potential for other mutations in this gene. In the case of S-gene target failure, there has been an inadvertent positive impact on diagnostic testing. Comparison of sequencing results to S-gene target failure results showed that S-gene target failure can be used as a proxy for detection of the B.1.1.7 variant [16]. In areas where B.1.1.7 is common, S-gene target failure has allowed the variant to be monitored at scale via diagnostic testing programmes, allowing it to be better characterized and appropriate control measures to be introduced. For example, in England, analysis of S-gene target failure was used to show that the frequency of the B.1.1.7 variant increased relative to other variants during a lockdown period, which should have limited opportunities for transmission [16]. This suggested that this variant was more transmissible than other variants. In Portugal, S-gene target failure has also been used to track the frequency and geographical dispersion of the variant [38]. Use of S-gene target failure as a proxy for B.1.1.7 only works well if the prevalence of B.1.1.7 is high, as other variants also contain the Δ69/70 mutation, and so also cause S-gene target failure. In SARS-CoV-2 variants 9 countries where there are several circulating variants with the Δ69/70 mutation, such as the U.S., the use of S-gene target failure may not be a suitable proxy for B.1.1.7. However, the presence of S-gene target failure can still be useful to select potential cases of interest, which can