Selectivity Window Test¶
Characterization test — instantiates Bycatch-Aware Selective Intervention Design
Sweeps the selector across its control variable to map where it separates target from non-target, locating the operating window in which selectivity holds and the edges where it collapses.
A Selectivity Window Test characterizes how selective a selector is across the full range of its control variable — dose, threshold, mesh size, aggressiveness — and finds the band within which it catches the target while sparing non-targets. Its defining move is treating selectivity as a curve, not a constant: it deliberately sweeps the setting from too-weak to too-strong and maps the resulting specificity, exposing both the window where the intervention works cleanly and the edges where target capture and bycatch become inseparable. It answers not "is this setting safe?" but "how wide is the safe band, and how close are we to its edge?"
Example¶
A drug-discovery team has a candidate antifungal that kills the pathogen but, because fungi are eukaryotes like us, risks harming human cells (the non-target). Rather than pick a dose and hope, they run a selectivity window test in vitro: they sweep concentration across, say, a ≈100-fold range and measure two curves at each step — pathogen kill and host-cell toxicity. The gap between the curves is the specificity profile; the concentrations where pathogen kill is high while host toxicity stays low are the selectivity window.
The test reports a window boundary — for instance, the candidate is only usable up to roughly 4× the effective dose before host toxicity climbs sharply — the ratio clinicians will later call a selectivity or therapeutic index. A narrow window flags the compound as hard to dose safely long before it ever reaches a patient; a wide one green-lights it for the next stage. Either way, the team now knows the band, not just a point.
How it works¶
- Sweep the control variable deliberately across its full plausible range, pushing past both edges of the useful setting rather than testing one nominal value.
- Measure target capture and non-target harm at each step, as paired curves rather than single points.
- Read the specificity profile as the separation between the curves, and the window as the band where target capture is high and bycatch stays below tolerance.
- Report the boundary and its margin — how much headroom exists before selectivity collapses — so operators know how tightly the setting has to be held.
Tuning parameters¶
- Sweep range and resolution — how far past the useful setting to push and how finely to step. Finer resolution locates the boundary precisely but multiplies test cost.
- Selectivity metric — how the target/non-target gap is scored (a ratio, a difference, the area between the curves). Different metrics move where the "boundary" falls.
- Test realism — a cheap bench or in-vitro proxy versus field or in-vivo conditions. A proxy often reports a wider window than the real setting delivers.
- Boundary criterion — how much bycatch defines the edge of the window; a stricter criterion narrows the reported window.
- Non-target coverage — one representative class versus several. More classes catch class-specific collapse but cost more to run.
When it helps, and when it misleads¶
Its strength is that it replaces "pick a setting and monitor" with a map of the whole trade-off, so operators learn not just a safe setting but how wide the safe band is — a wide window tolerates sloppy operation, a narrow one demands tight control or a different method entirely. It also turns selectivity from an anecdote into a comparable number across candidate selectors.
Its results are only as valid as the test conditions: a window measured on a bench proxy or a single non-target class can be far wider than the messy field delivers, lending false confidence right where it is most dangerous. The narrowness of a window is a property of the selector, not the test — a test cannot widen a fundamentally blunt instrument — and reporting a comfortable window from unrealistic conditions is the classic misuse (a roomy in-vitro window that collapses in vivo). The discipline that guards against it is to test against the real non-target classes under conditions as close to deployment as affordable, and to read a narrow window as a signal to retune or substitute, not merely to operate more carefully. The ratio it reports is the same idea pharmacology calls a therapeutic or selectivity index.[n1]
How it implements the components¶
selector_specificity_profile— the test's primary product: the paired target-capture and non-target-harm curves traced across the control variable.selectivity_window_boundary— the operating band it extracts from that profile, reported with the margin to each edge.
It measures the window but does not act on it: iterating the selector to widen a narrow window belongs to Selector Retuning Cycle, and swapping to a different method when the window is hopeless belongs to that cycle's method-substitution branch. Deciding what counts as on-target in the first place comes from Non-Target Impact Pre-Mortem.
Related¶
- Instantiates: Bycatch-Aware Selective Intervention Design — the window test supplies the specificity map the design uses to set, hold, or abandon a selector setting.
- Consumes: Non-Target Impact Pre-Mortem supplies the target and non-target definitions the test measures the separation between.
- Sibling mechanisms: Selector Retuning Cycle · Non-Target Impact Pre-Mortem · Non-Target Sentinel Sampling · Success Metric Reweighting · Bycatch Rate Dashboard · Bycatch Tolerance Stop Rule · Compensation and Restoration Trigger · Escape Hatch or Release Protocol · False-Capture Audit · Negative Filter or Exclusion Device
Editorial Notes¶
Form Classification¶
Form family: Experiment, Test & Rehearsal
Rationale: Selectivity Window Test operates as an active test, trial, simulation, drill, or rehearsal that generates evidence through a deliberate attempt or perturbation because it sweeps the selector across its control variable to map where it separates target from non-target, locating the operating window in which selectivity holds and the edges where it collapses.
Independent corroboration: The frozen evidence defines Selectivity Window Test as 'Sweeps the selector across its control variable to map where it separates target from non-target, locating the operating window in which selectivity holds and the edges where it collapses', so its operative form is Experiment, Test & Rehearsal.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Engineering & Design
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Testing the range over which a selector responds to targets while rejecting near alternatives is an engineering characterization of selectivity. IUPAC defines selectivity as the extent to which a method determines one analyte without interference, with statistics supplying the test design.
Related originating lineages:
- Biology & Ecology — Biological and ecological research supplies a parallel or contributing lineage for the mechanism's defining operation: sweeps the selector across its control variable to map where it separates target from non-target, locating the operating window in which selectivity holds and the edges where it….
- Chemistry & Materials Science — Process and separation chemistry routinely characterize selective temperature, concentration, or affinity windows.
- Data Science & Analytics — Data science, analytics, and operational monitoring supplies a parallel or contributing lineage for the mechanism's defining operation: sweeps the selector across its control variable to map where it separates target from non-target, locating the operating window in which selectivity holds and the edges where it….
- Mathematics — Mathematical modeling, proof, and abstract-structure practice supplies a parallel or contributing lineage for the mechanism's defining operation: sweeps the selector across its control variable to map where it separates target from non-target, locating the operating window in which selectivity holds and the edges where it….
- Pharmacology & Toxicology — pharmacology_toxicology contributes dose response, analyte interference, safety windows, and exposure to this mechanism's defining operation—Sweeps the selector across its control variable to map where it separates target from non-target, locating the operating window in which selectivity holds and the edges where it collapses—without displacing the selected primary historical lineage.
- Physics — Instrumental discrimination often depends on resolving regions across a tuned physical variable.
- Statistics & Experimental Design — Designed sweeps and response surfaces estimate the window and uncertainty at its boundaries.
Review resolution: The blind reviewers disagree on primary lineage (engineering_design versus statistics_experimental_design). Authoritative or primary research supports engineering_design as the best historical origin: Testing the range over which a selector responds to targets while rejecting near alternatives is an engineering characterization of selectivity. IUPAC defines selectivity as the extent to which a method determines one analyte without interference, with statistics supplying the test design. The cited IUPAC Gold Book, Selectivity directly supports the mechanism's defining operation. All independently supported contributing domains are retained without an arbitrary cap. origin_mode=cross_disciplinary_synthesis records the lineage relationship, while domain_reach=multi_domain records later applicability separately from provenance.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Researched adjudication after independent review; high confidence.
Sources consulted:
Notes¶
[n1] The therapeutic index (or, for a specific target/off-target pair, selectivity index) is the ratio between the dose that harms the non-target and the dose that acts on the target — pharmacology's standard measure of how wide a selectivity window a compound offers. A large index means a forgiving window; a small one means the effective and harmful doses sit dangerously close. ↩