A DNA design loop retargets SARS-CoV-2 detectors in days, not months

A research group has compressed the redesign of DNA aptamers, short engineered strands that grab a specific molecular target, against SARS-CoV-2 variants from many months to days. In a single run, the platform profiled 11,792 candidate designs against three spike proteins, pulled out a Delta-binding mutant with 4x the original affinity, and turned an Omicron-binding mutant from undetectable into a nanomolar grip. The work, published this month in ACS Nano, positions the screening pipeline as a generalizable design loop rather than a one-off result.

The reference point matters. Spike-targeting monoclonal antibodies typically took 10 to 12 months to develop after the original SARS-CoV-2 outbreak, and individual candidates have lost potency as the virus evolved. Aptamers have been explored as an alternative recognition chemistry since the 1990s, but their discovery workflow has been the bottleneck. The new platform attacks that bottleneck directly.

The pipeline starts with a 40-nucleotide aptamer originally selected against wild-type spike. A random-rational hybrid library diversifies that seed, swapping bases at positions the team expected to matter and randomizing the rest. The library is then profiled on a repurposed MiSeq sequencer adapted as a high-throughput binding screen. Each read reports a candidate's interaction with a chosen spike protein, so a single run can map thousands of designs against multiple targets in parallel. The screen takes days from pool amplification to data analysis, the authors report.

The output is a structured binding map. From that map, the team extracted a Delta-binding mutant with roughly fourfold improved affinity over the wild-type parent. They also extracted an Omicron-binding mutant that converted previously undetectable binding into nanomolar affinity, which is the range most diagnostic probes aim for. A third mutant, engineered for wild-type selectivity, dropped its Delta affinity substantially; the team also identified the specific bases that drive variant discrimination. Molecular dynamics simulations were then layered on the binding data to rationalize which substitutions accounted for the variant-specific behavior.

To show the candidates work outside the sequencer, the team built fluorescent strand-displacement sensors from the wild-type-selective and Omicron-selective mutants. The sensors distinguished wild-type, Delta, and Omicron spike proteins in solution, which is the basic shape of a diagnostic readout even though no clinical test was run.

The limits are explicit in the paper. The screen is a single research group's demonstration. The Delta and Omicron validation is retrospective, working against lineages that have already circulated rather than a newly emerging variant. No clinical samples, no real-world diagnostic deployment, and no production-scale throughput or unit-economics data are reported. The broader claim that the platform can rapidly adapt to other pathogens is an author extrapolation, not a result in this paper. SARS-CoV-2 itself is no longer an acute 2026 priority. The contribution to track is whether the workflow reproduces on a different target.

The watch item is generalization. The screening chemistry, the random-rational diversification strategy, and the MiSeq-based readout are platform-level, not virus-specific. If independent groups reproduce the design-loop speed on a different target, the result is a faster tool for keeping molecular recognition current with whatever pathogen comes next. Until then, the platform is a research-stage workflow with concrete Delta and Omicron proof points and a sensor that works in a beaker.