tide_quantify_indels
Read-only
Quantify a CRISPR edit from a pair of Sanger traces by TIDE (Brinkman et al.) — the abundance of each indel SIZE in the sequenced population, with a p-value per size class, the total editing efficiency, the goodness of fit, and the base composition of the +1 insertion.
Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| guide | Yes | the guide spacer, RNA accepted. TIDE decomposes around ONE break, so this is a single guide rather than a list — it is located in the control on either strand and the break is placed three bases 5-prime of the PAM. | |
| _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. | |
| edited | Yes | the EDITED trace, by file name or id. | |
| control | Yes | the UNEDITED trace, by file name or id — every indel class is the control trace shifted, so this read IS the model. | |
| seq_end | No | last base of the READ to use, clamped down to the shorter of the two reads if it overruns. In bp. | |
| _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. | |
| max_shift | No | largest indel size modeled, in both directions. One basis function is fitted per size from -max_shift to +max_shift, so this bounds what can be reported at all. In bp. | |
| seq_start | No | first base of the READ to use. The start of a Sanger read is poor quality; upstream's own web app starts at 100. In bp. | |
| window_end | No | last base of the decomposed region. 0 = as far as the shorter read allows. In bp. Advanced. | |
| p_threshold | No | an indel class is called PRESENT below this p-value. A 0% class and a non-significant class are different answers, which is why this is a knob rather than a constant. Advanced. | |
| window_start | No | first base of the decomposed region. 0 asks the engine to place the window after the break itself, which is what upstream does when given no boundaries. In bp. Advanced. |
Output Schema
| Name | Required | Description | Default |
|---|---|---|---|
| indels | Yes | every modeled size from -max_shift to +max_shift with its percentage, p-value and whether it cleared the threshold — INCLUDING the classes that did not, so the spectrum has a shape and not only peaks | |
| report | Yes | the whole result, including the window corrections upstream raises as warnings | |
| rSquared | Yes | goodness of fit of the whole decomposition. Every percentage below is already scaled by it, so a low fit shrinks the classes rather than flagging them | |
| breaksite | Yes | the base the break was placed at, three bases 5-prime of the PAM. Every window in the result is positioned against it | |
| efficiency | Yes | total editing efficiency: the R-squared-scaled signal in every non-zero indel class, 0-100. Not the same quantity as an ICE score — wire them into a comparison rather than treating either as the number | |
| aberrantSignal | Yes | the four means TIDE's own quality checklist is read from: control and sample, before and after the break. A sample already aberrant BEFORE the break is a sequencing problem rather than an edit, and this is the pair of numbers that says so | |
| insertionBases | Yes | base composition of the +1 insertion as percentages in A, C, G, T order — the readout ICE cannot produce. Null when there is no significant +1 class, which is the common case and not an error; `insertionNote` then says why in upstream’s own words | |
| significantIndels | Yes | the classes that cleared p_threshold, excluding the unmodified class — what a gate should read |