What Can In Silico Inclusivity Show for a Respiratory Panel?

What a sequence match supports

A molecular respiratory panel team finds that its primers and probes match 98% of records in a public sequence database. In this hypothetical example, 98% describes the records that met the team’s matching rule. It cannot support a statement that the panel detects 98% of circulating variants. The denominator is a set of sequence records, and the analysis contains no observations of how the finished assay performed.

For a US submission or labeling review, use the sequence screen to document which binding sites aligned, which mismatches appeared, and which variants warrant wet testing. Record bench results by strain or material, specimen matrix, concentration, assay configuration, and replicate outcome. A positive finding supports analytical reactivity under those study conditions. A claim about current circulation needs dated surveillance evidence for the relevant population and geography.

What the in silico screen can establish

For each claimed target, align the assay’s forward primer, reverse primer, and probe against the corresponding sequence regions. Define what counts as a match for each oligonucleotide and for the target as a whole. An assay that uses several targets may tolerate a change in one binding site, depending on how the signals are combined into a reported result. Record the individual calls and the overall match rule so a reviewer can see which result might be affected.

A summary can report how many evaluable records met that rule. Retain the individual mismatch patterns underneath it. A row that lists two substitutions in a forward primer, with their positions relative to its 3′ end, gives the lab a specific challenge to investigate. A sequence with an ambiguous base or a gap at the binding site belongs in the indeterminate or excluded count according to a documented rule.

What needs a laboratory measurement

An alignment evaluates the sequence relationship without exercising specimen preparation, extraction, amplification, or the instrument’s result algorithm. In a multiplex panel, distinguish loss of an individual signal from a change in the result delivered to the user. For a strain that reacts at a higher concentration, record the tested levels and the result at each level. Confirming a strain-specific limit of detection (LoD) calls for a study designed to establish that limit. See the guide to strain and matrix coverage for respiratory panel LoD.

A synthetic template can help isolate the effect of a particular mismatch. If it enters the workflow after sample preparation, record the bypassed steps. An isolate or inactivated material in a representative matrix can answer a broader analytical question when it goes through the intended workflow. Those contrived results still describe a controlled study. Report clinical performance from appropriate clinical evidence.

Four fields for a reproducible sequence record

• Assay configuration. Identify the target and reported result, assay or oligonucleotide revision, and reference accession and coordinates for each binding site. Keep the strand and coordinate convention explicit so a reverse-primer mismatch can be located unambiguously.

• Repository and query. Record the repository name, release or retrieval date, exact query, and collection period and geographic filters where used. Distinguish collection dates from deposition dates. Retain accession versions or a reproducible list of the records analyzed.

• Curation and alignment rules. Record retrieved, excluded and evaluable counts, including handling of duplicates, incomplete binding sites, ambiguous bases and gaps. Preserve the tool version, alignment parameters, match definition, and positional and substitution or indel calls for each oligonucleotide.

• Disposition. Record why the observed pattern was selected for wet testing, deferred pending material, or scheduled for further monitoring. Assign an owner and recheck date. If a model was used to prioritize testing, record its assumptions and keep its output with the original screen.

Keep the denominator beside the match percentage. If you deduplicate identical genomes, make clear whether the percentage refers to deposited records or unique sequences. Show exclusions by reason so that a high match percentage cannot hide poor coverage of the binding regions. When a query or tool changes, document the change before comparing successive snapshots. Otherwise an apparent rise in mismatches may reflect the new counting method.

What the LIAISON PLEX studies showed

FDA’s 510(k) decision summary for the LIAISON PLEX Respiratory Flex Assay, K233410, describes a combination of sequence analysis and laboratory testing. The sponsor prepared contrived specimens in simulated nasopharyngeal swab matrix, with many strains initially tested in triplicate at approximately three times the assay LoD. The study retested strains at the same concentration after a missed detection, then used higher levels when the same-level retest did not achieve detection in every replicate. FDA decision summary, analytical reactivity section.

The sequence analysis used influenza records uploaded to GISAID from September 1, 2015, to July 7, 2023. For the influenza A H1 target, 53,778 of 54,364 sequences (98.9%) met the study’s greater-than-90% oligonucleotide identity criterion. That is a prediction for an archived sequence set. The test runs below have their own materials, concentrations, and denominators. K233410, Table 24.

For influenza A strain A/Hawaii/66/2019 X-345A, the H1 subtype was detected in two of three initial replicates and three of three on retest, or five of six pooled. The influenza A matrix gene target was detected in all six replicates. For Brisbane/02/18, the first tested level with detection in every replicate was 9× the reference LoD. A Bordetella holmesii strain, CIP 104396, reached that study endpoint at 27× after misses at lower levels. K233410, Tables 11 and 18.

The Hawaii result illustrates why the ledger needs both the broad influenza A result and the subtype call, plus the original and repeat counts. Here “matrix” names an influenza A gene target in the table. The specimen matrix is the simulated swab material. The 9× and 27× figures describe study concentrations with detection in every replicate tested. They supply neither a newly confirmed LoD nor a recommended test concentration for another panel.

What the mismatch study found

Otwell and colleagues tested 16 SARS-CoV-2 PCR assays with synthetic templates. Many mismatched templates retained amplification without a large loss of performance. The researchers also reported false negatives for an N501Y assay in retrospective testing of Omicron specimens. Their screen flagged a potential failure only when a primer or probe exceeded 10% mismatch. Two mismatches in a 20-base forward primer did not cross that cutoff. Follow-up template experiments implicated those two changes, while a reverse-primer change near the 3′ end had little effect in that assay. Otwell et al., 2025, Table 2 and accompanying results.

Use position, type, and combinations of changes to choose what to investigate. A count threshold is a screening choice whose performance can vary by target and assay. Preserve the original prediction even when later bench testing shows reactivity. If a screen shows little concern and the lab observes missed detections, preserve both findings and investigate the discrepancy under the actual assay conditions.

Evidence for a current circulation claim

A repository snapshot shows what was deposited and selected for analysis. A historical strain panel shows which materials were tested and how they performed at the time. To say a lineage is circulating, record the surveillance source, collection period, geography, and limits of the surveillance sample. Then link the lineage to the sequence and assay evidence separately. This keeps a recent upload of an old isolate from being treated as recent circulation.

FDA’s November 10, 2026 final guidance addresses certain emerging-pathogen IVDs during an applicable Section 564 declaration. For molecular inclusivity, it recommends screening primers and probes in silico and supplementing the analysis with wet testing of currently circulating variants when available, using clinical isolates or inactivated material spiked into clinical matrix at or near the assay LoD. It generally recommends full sequence identity and recommends assessing the potential performance impact when a significant number of published sequences contain mismatches. It also recommends monthly inclusivity evaluation and a risk analysis if testing shows a greater than threefold reduction in analytical sensitivity between pathogens with and without the mutation. These are nonbinding recommendations for that emergency context. FDA emerging-pathogen validation guidance, Analytical Validation Testing, Inclusivity.

The US Class II special controls in 21 CFR 866.3981 apply to the device type defined there for detecting SARS-CoV-2 and other microbial agents in human respiratory specimens with a multi-target test. They require documentation of a continuous monitoring protocol for mutations and novel strains and documentation of its results and findings. Periodic in silico analysis and literature review are examples in the regulation. The labeling must communicate the risk of incorrect results from target-sequence variants and describe relevant performance, including inclusivity as applicable, for each claimed specimen type. 21 CFR 866.3981(b)(3)–(4).

FDA’s 2009 guidance for respiratory viral panel multiplex nucleic acid assays under 21 CFR 866.3980 describes the risk of missed detection when new variants or subtypes affect primer or probe reactivity. It discusses multiple probes as one design control where applicable and describes analytical reactivity as part of the performance evidence. FDA respiratory viral panel special-controls guidance.

Build a record that links the screen and wet test

Create a ledger row for each target and variant pattern on a defined assay version. Link it to the underlying sequence report and any wet-test run. Record material identity, matrix, workflow, concentrations in units and relative to the reference LoD, the predefined interpretation rule, and the number detected out of those tested. Preserve invalid runs and the reason for any retest. The following entries are hypothetical and illustrate recordkeeping.

If the lab tests a surrogate, identify the accession or mismatch pattern it represents and how the surrogate differs from the intended specimen. If a mismatch spans more than one target, record which target or channel dropped out and which result was ultimately reported. Keep a specific disposition for each open question so a favorable result for one target cannot close another target’s investigation.

Write the claim and limitation together

A reviewable statement links each observation to its method and scope. For a sequence result, use wording such as “In the [repository] snapshot retrieved on [date], [number] of [denominator] evaluable records met the defined primer and probe matching criteria.” Follow with a distinct bench statement such as “In [matrix], [number] of [denominator] valid replicates of [material] were detected at [concentration] on [assay configuration].” State what remains untested, including variants or matrices outside the study, next to the corresponding result.

If a summary combines all respiratory targets, check whether it hides a weaker result for one strain or for a subtype. Report invalid tests alongside the valid denominator, and retain the first run when summarizing a retest. Use the configured intended result in the evidence record. A detected organism-level target can coexist with a missed subtype, as the K233410 example shows.

Inclusivity asks how the assay reacts to diversity within an intended target. Cross-reactivity asks whether other targets or organisms produce unintended signals. The study question determines the material and challenge level. See the respiratory panel study comparison.

When to reopen the evidence record

• New sequence or surveillance signal. Reassess when a new pattern appears at a binding site or when dated surveillance changes the relevance of a lineage. Keep the previous snapshot so the reviewer can trace the change.

• New assay or specimen configuration. Review whether changes to oligonucleotides, extraction, matrix, instrument, or interpretation software affect which prior evidence still applies. Record the rationale before carrying a result into revised labeling.

• Unexpected wet result or changed claim. Keep misses and invalids visible. Route a suspected performance impact through the applicable risk and regulatory process. Revisit the evidence whenever wording changes from a sequence observation to a claim about analytical detection or circulation.

For devices within § 866.3981, the monitoring procedure should account for the rule’s reporting provisions. Results reasonably interpreted as showing that novel respiratory strains or isolates affect stated expected performance must be sent to FDA immediately. The section also specifies submission of evaluations within 48 hours when FDA requests them. 21 CFR 866.3981(b)(4)(ii).

What to hand the next reviewer

Hand over the versioned ledger, sequence accession list and curation rules, and the wet-test reports with material, concentration, matrix and replicate details. Include the current claim and limitation alongside any open discrepancy, assigned owner and next review date. When “currently circulating” appears in a draft, the handoff should identify the dated surveillance record behind that wording and the assay evidence tied to it.