Introduction
Multiplex PCR relies on gene-specific primers to enrich for viral sequences, and this makes it useful for generating viral genomes sequencing from clinical samples with a complex background. This also, however, makes it susceptible to variation in the viral genome reducing the binding efficiency of primers. The Omicron variant has a very large number of mutations especially in the Spike protein. Some of these are shared with existing variants which will not cause a problem but many of them are novel or have been seen at low frequency. The current ARTIC V4 primer set is designed around lineage-defining mutations in 8 lineages (B.1.1.7, B.1.351, B.1.429, B.1.525, B.1.617.1, B.1.617.2 and P.1) but since it’s release in June 2021 the emergence of Delta sub-clades e.g. AY4.2 which accounts for an increasing number of cases in the U.K. and the Omicron variant have caused collisions with primers in the scheme, reducing the likelihood of it yielding complete genome sequences which is what we strive to achieve.
Problem observed
The following primers are affected by mutations in the Omicron variant. It is difficult to predict the effect of these mutations on amplification efficiency although in the past we have observed that 3’ mutations are most likely to result in a drop-out. We have now analysed data available via CLIMB for Omicron variant and we can confirm that amplicons 76, 79 and 90 result in near complete dropouts.
| Site | Primer affected |
|---|---|
| 2832 | SARS-CoV-2_10_LEFT |
| 8393 | SARS-CoV-2_27_RIGHT |
| 22673 | SARS-CoV-2_76_LEFT |
| 22674 | SARS-CoV-2_76_LEFT |
| 23040 | SARS-CoV-2_76_RIGHT |
| 23048 | SARS-CoV-2_76_RIGHT |
| 23055 | SARS-CoV-2_76_RIGHT |
| 23948 | SARS-CoV-2_79_RIGHT |
| 26270 | SARS-CoV-2_88_LEFT |
| 26577 | SARS-CoV-2_89_LEFT |
| 27259 | SARS-CoV-2_90_RIGHT |
Investigating these regions in more detail we can see 76_LEFT being affected by two adjacent SNPs both 3’ and 76_RIGHT affected by three SNPs. 90_RIGHT is affected by a 3’ SNP (below).
Solution
V4 was released in June (see earlier post for details) to handle a number of problematic mutations in Beta and Delta variants. At the time we thought that designing a new scheme from scratch using the latest version of primalscheme and validating it would be a better solution than trying to update it. This time we are proposing to update the V4 to V4.1 using a patch consisting of 11 new primers which have been designed to be either used as a spike-in to existing V4 pools or as a replacement for the respective primers. The replacement 23_RIGHT primer in the V4.1 patch is to address a mutation that occurs in AY.4.2 which accounts for an increasing proportion of cases in the U.K. and requires only a small change so we are also including it in this release. We hope these changes are sufficient to restore the performance for Omicron but they have been designed in silico and not yet experimentally validated. We are releasing the sequences ahead of testing them to allow people to place orders for oligos while testing is ongoing.
Methods
As mentioned in the previous post our preference is now to use N-masked reference genomes rather than unaligned genomes in multi-FASTA format. This has the advantage of greater granularity when designing primers between mutations. For this update the lineage defining mutations from B.1.1.529 and AY.4.2 we added to the input file used to design V4 (B.1.1.7, B.1.351, B.1.429, B.1.525, B.1.617.1, B.1.617.2 and P.1). A new option to repair primer schemes was added to primalscheme. This can be accessed using the --repair flag when providing an updated FASTA input file and an existing BED file. In this mode the program checks if the sequences in the BED file still exist in the FASTA file and if not attempts to repair them within the design constraints of the scheme. This was able to successfully repair the V4 scheme for Omicron without having to completely redesign the scheme. We hope this will allow us to update schemes faster and more cheaply with the emergence of future variants. Massive thanks to Andy Smith the co-developer of primalscheme for getting this working in under two days!
| MN908947.3 | 2780 | 2813 | SARS-CoV-2_10_LEFT_alt1 | 2 | + | TGAATATCACTTTTGAACTTGATGAAAGGATTG |
| MN908947.3 | 3156 | 3177 | SARS-CoV-2_10_RIGHT_alt1 | 2 | - | GGTTGAAGAGCAGCAGAAGTG |
| MN908947.3 | 7127 | 7156 | SARS-CoV-2_23_RIGHT_alt1 | 1 | - | AGAATCTAAACCACTAAGACAAACACTAC |
| MN908947.3 | 8367 | 8392 | SARS-CoV-2_27_RIGHT_alt1 | 1 | - | AATGTTGTGACTTTTTGCTACCTGC |
| MN908947.3 | 22742 | 22774 | SARS-CoV-2_76_LEFT_alt1 | 2 | + | ATGTCTATGCAGATTCATTTGTAATTAGAGGT |
| MN908947.3 | 23120 | 23141 | SARS-CoV-2_76_RIGHT_alt1 | 2 | - | GTCCACAAACAGTTGCTGGTG |
| MN908947.3 | 23914 | 23944 | SARS-CoV-2_79_RIGHT_alt1 | 1 | - | AATTGGTGGTGTTTTGTAAATTTGTTTGAC |
| MN908947.3 | 26242 | 26268 | SARS-CoV-2_88_LEFT_alt1 | 2 | + | TTATGTACTCATTCGTTTCGGAAGAG |
| MN908947.3 | 26592 | 26621 | SARS-CoV-2_89_LEFT_alt1 | 1 | + | TAGGTTTCCTATTCCTTACATGGATTTGT |
| MN908947.3 | 26966 | 26991 | SARS-CoV-2_89_RIGHT_alt1 | 1 | - | CTAGATGGTGTCCAGCAATACGAAG |
| MN908947.3 | 27218 | 27251 | SARS-CoV-2_90_RIGHT_alt1 | 2 | - | ATTAGTAATATCTCTGCTATAGTAACCTGAAAG |
To use these either repool V4 using the pooling volumes given in the README or pool alts into two spike-in pools then add 0.2uL per 4uL (10 uM) into the respective mastermix to give a final per oligo concentration of ~15nM.
Thanks to the WTSI and Naomi Parks for generating the data and reporting on V4 performance.