The Twenty-Eight Day Rule Is Over: NGS Enters Mandated Release Testing

The unglamorous half of the sequencing market

Ask what next-generation sequencing is for and most answers describe discovery: single-cell atlases, tumour profiling, population cohorts, the long-read push into structural variants. There is a second use, considerably less discussed and considerably more reliable, and it is quality control. Sequencing is now an accepted method for proving that a biological product does not contain a virus it should not contain, and that use is written into a harmonised international guideline rather than into a purchasing budget.

That distinction – discretionary versus mandated – is the most underrated fact about the sequencing market. Discovery demand rises and falls with funding cycles. Release-testing demand rises with the number of batches manufactured, because a batch cannot be released until the test is done. One of those lines is cyclical; the other is attached to industrial output.

What ICH Q5A(R2) actually changed

The International Council for Harmonisation adopted Q5A(R2) on viral safety evaluation of biotechnology products derived from cell lines of human or animal origin on 1 November 2023, and it did something the previous framework could not. It opened the door to next-generation sequencing as an alternative to the classical adventitious-agent testing battery.

Three changes matter commercially. First, non-targeted next-generation sequencing is now accepted for testing unprocessed bulk, which is where the indicator-cell assay and the in vivo animal assay traditionally sat. Those are slow, insensitive to some virus families and increasingly difficult to resource. Second, the 28-day cell-culture period can be reduced to 14 days where a risk assessment supports it – and reducing two weeks from a release pathway is an operational gain, not a scientific one. Third, nucleic-acid testing and targeted sequencing are accepted for specific virus families where the classical methods are weak.

The guideline does not say sequencing is better. It says sequencing is acceptable when it is validated and when the bioinformatics pipeline is qualified – which is a harder condition than buying an instrument, and it is where the money in this market actually goes.

Why this is a bioinformatics market wearing an instrument’s clothes

A release test has one job: to produce a result that a regulator will accept and that a manufacturer can defend in an inspection. That is a different product from a research sequencer. It requires a reference virus database, a documented pipeline, defined thresholds for what counts as a detected sequence, a validated limit of detection for each virus family of interest, and a change-control process for every software release.

The instrument is the least differentiated part of that stack. Several platforms can generate the reads; the qualification of the analytical pipeline is proprietary, expensive and slow. This is precisely the mechanism described in the NGS market report: as sequencing moves from a research tool into a regulated tool, value migrates away from the box and toward the validated method sitting on top of it. It is the same reason the sequencer stopped being the product in the clinical market.

The commercial footprint of that migration is visible in instrument-maker results. Illumina’s 2025 revenue of USD 4.34 billion was flat overall, with the company’s own ex-China figure of USD 4.10 billion growing 2 percent – the difference being a single geography, not a category. Oxford Nanopore, by contrast, grew 24.2 percent at constant currency to GBP 223.9 million, and the composition of that growth is the interesting part: clinical revenue up 59.9 percent and bioPharma up 30.4 percent, against research at 15.1 percent. Those two faster lines are the ones that sit closest to regulated and industrial testing rather than to grant-funded discovery.

The upstream market that needs the test

Adventitious-agent testing exists because of a manufacturing process, and the process that generates the most testing per unit of product is transient transfection. A typical AAV run using PEI-mediated transfection of HEK293 cells completes in about a week and can produce more than 1 x 10^14 genome copies at over 90 percent capsid purity. Every one of those runs needs a safety package, and the number of runs is what scales the testing market – not the litres of capacity.

That is why the mammalian transient protein expression market, forecast from USD 1.48 billion to USD 3.42 billion at 11.04 percent, is a leading indicator for sequencing-based safety testing rather than an unrelated niche. The two markets share a driver: the number of experiments and batches started. Transient processes are non-clonal and unbanked by construction, so each run is a discrete production event with its own testing obligation, as the companion piece on why transient expression has a structural ceiling sets out.

What a buyer is actually purchasing

Read the procurement decision from the manufacturer’s side and the logic becomes clear. The classical battery costs time, animals and cell-culture capacity, and it is incompletely informative: an in vivo assay will not detect a virus that does not replicate in the species used. A validated sequencing method costs a pipeline qualification and then, per test, less time and better breadth.

The real cost is the qualification. Assembling the reference database, establishing limits of detection across virus families, and defending the threshold-setting logic consumes specialist bioinformatics capacity that most manufacturers do not staff. That is why the fastest-moving adopters are not the largest manufacturers but the ones with in-house analytical development and a regulatory affairs team comfortable with computational methods.

It is also why the service layer matters. A contract testing organisation that already holds a qualified sequencing pipeline can sell a validated result rather than an instrument, and that is a considerably better business than reselling hardware into a market where the hardware is converging. The diagnostic testing report describes the same value migration in the clinical laboratory, where the annuity sits in the assay menu rather than the analyser.

The constraint nobody has solved

The limiting factor is not read length, throughput or cost per gigabase. It is the absence of a common analytical reference. Two laboratories running the same sample through different pipelines can reach different conclusions about a marginal signal, and there is no agreed benchmark against which either can be judged. Until there is, every manufacturer’s validation stands alone, every inspection is a fresh argument, and adoption proceeds at the pace of individual regulatory comfort rather than at the pace the technology allows.

That is a solvable problem and it is being solved in public: guideline revision, method standardisation and the slow accumulation of regulatory precedent. It is simply slower than the instrument roadmap, and it is the reason a mandatory use case can still grow gently. The companies positioned for it are not the ones with the best sequencer. They are the ones whose pipeline qualification a regulator has already seen.