bootstrap_cq
Resampling error bars on ONE reaction's Cq. A Cq is read off a fitted model, so its uncertainty is how tightly the observations pin that model down - which propagating a measured spread cannot answer.
Tool call template
Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.
{
"id": 1,
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"arguments": {
"curves": "<curves>"
},
"name": "bootstrap_cq"
}
}Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| drop | No | jackknife only: observations left out of each refit. 1 is the delete-one jackknife; more is a delete-d jackknife, which is less sensitive to a single influential cycle. Only applies when scheme is jack. | |
| seed | No | R's `set.seed` value, so the interval is reproducible and citable. Advanced. | |
| model | No | the sigmoid fitted once and REFITTED on every iteration. It is not re-chosen per resample: Cqs fitted under different models are not comparable, so a mixture over models would make the quantiles mean nothing. Values: l4 (l4 · 4-param log-logistic), l5 (l5 · + asymmetry), l6 (l6 · + linear baseline drift), l7 (l7 · + quadratic baseline), b4 (b4 · 4-param logistic (raw scale)), b5 (b5 · + asymmetry), b6 (b6 · + linear baseline drift), b7 (b7 · + quadratic baseline). | l4 |
| _files | No | The project: record references shaped like _subject. Their records become library documents a tool names in sequences/parts/document; a trace or alignment is found by file name. | |
| curves | Yes | the amplification table: a header row, first column the cycle number and every other column one reaction's fluorescence. The same parser `call_cq` and the Curve->Cq tab use. | |
| scheme | No | `boot` resamples the fit's residuals (upstream's default); `jack` leaves observations out. The bootstrap answers "how would this Cq move under the same noise?", the jackknife "how much does any single cycle carry it?". | boot |
| _subject | No | The record to act on: exactly one of sequence, accession (+source), sample (an id from list_samples) or file ({name, text} or {name, base64}); record_index picks one record of several. | |
| reaction | No | which reaction to resample, by its column name. Blank takes the first column — this is a per-reaction run by design, because at the default 200 iterations a 47-reaction plate would be 9,400 refits nobody asked for. | |
| cq_method | No | which Cq definition the resampled efficiency readouts report. Values: cpD2 (cpD2 · second-derivative max), cpD1 (cpD1 · first-derivative max), maxE (maxE · cycle of maximum efficiency), expR (expR · exponential region), Cy0 (Cy0 · tangent x-intercept (Guescini)), CQ (CQ · 20 % of the Cq fluorescence), maxRatio (maxRatio · spline, model-free (Shain)). | cpD2 |
| conf_level | No | the level the reported bounds are read at — 0.95 puts them at the 2.5 % and 97.5 % quantiles of the iterations. | |
| iterations | No | each iteration is a full nonlinear refit, so this is the run's whole cost. 100 is upstream's default and puts a 95 % bound on a single order statistic, which is why both doors open on 200. | |
| include_readouts | No | also resample every efficiency readout — the Cq definitions, the efficiency, the initial fluorescence — not just the model coefficients. On, because an interval on the Cq is the point; off is faster and returns coefficients alone. Advanced. |
Output Schema
| Name | Required | Description | Default |
|---|---|---|---|
| rss | Yes | residual sum of squares across the iterations | |
| rmse | Yes | residual standard error across the iterations | |
| model | Yes | the model refitted on every iteration | |
| reaction | Yes | the reaction actually resampled, named — "the first column" is not a citation | |
| readouts | Yes | one row per resampled readout — Cq under each definition, efficiency, initial fluorescence — with mean, SD, median and the interval quantiles. Empty when `include_readouts` is off | |
| confLevel | Yes | the level the bounds above are read at | |
| converged | Yes | iterations whose refit converged — the ONLY ones any number here is computed from. Well below `iterations` is a finding about the curve, not a warning to swallow | |
| iterations | Yes | iterations requested | |
| coefficients | Yes | one row per model coefficient, in the model's own parameter order, same fields |