Electropherogram

Every dye channel of a fragment run on one axis in bases, with the size standard's own fragments drawn on it — the picture the peak table is read from.

What it doesCopy Anchor Link

The instrument records a fragment run in scans: how far along the capillary each detection was. A scan is not a length, and a plot with a scan axis cannot be compared with anything — not with the same sample run on another day, not with a published allele size, not with the run in the next well.

So this canvas does not draw scans. Every sample of every channel is placed through the same sizing curve the peak table uses, so the axis is in bases and the peak you point at is the row you read. The dashed verticals are the size standard's own fragments: the points the curve was fitted through, drawn on the thing they measured, so you can see where the ruler has marks and where it is interpolating between them.

Each trace is drawn in an approximation of its dye's own emission colour, which is what every instrument's software does and the only thing that makes four or five overlapping channels legible. A dye the build does not recognise by name falls back to a colour-blind-safe palette rather than disappearing — real files carry typos in that field, and one of the ladders shipped here has two.

What this computesCopy Anchor Link

The axis. Every sample of every dye channel is placed in bases through the run's own size-standard sizing curve — the same curve the peak table is read from, re-fitted here from the same anchors rather than carried across as a second copy, so the picture and the row can never disagree. Five sizing models are offered, and the plot redraws in whichever one is selected: Local Southern (the reciprocal-fit default), Global Southern, least squares of second and of third order, and a natural cubic spline. The axis span is the size standard's own supported range intersected with the analysis window, so the plot never extrapolates past the last ladder fragment.

The ruler, drawn on the thing it ruled. The size standard's fragments appear as dashed verticals at the (scan, base-pair) anchors the curve was fitted through, which is what makes interpolation gaps visible: a stretch of axis with no dashed line is a stretch where the sizing is least well supported, and every size in it inherits that.

Signal. One path per dye channel of analyzed (instrument colour-corrected) signal, each in an approximation of its dye's emission colour, with a colour-vision-safe fallback for a dye name the build does not recognise. Channel scale is the knob: Shared puts every channel on one ceiling — the honest picture of relative signal — and Per channel scales each to its own maximum, which is what makes the size standard readable beside taller allele peaks. Both axes are labelled with their own numbers, and the per-channel view says on the plot that the ceiling differs per channel rather than printing one channel's counts over all of them.

Marks and readouts on the picture itself. Detector saturation (off-scale) is marked per peak above the trace, because a saturated peak's height is a floor rather than a measurement while its size is still read normally. The size-standard channel is named in the legend and can be hidden. Pointing anywhere reports the base-pair position under the cursor, and pointing at a peak reports its dye, size in bases to two decimals, and height; clicking selects that peak's row in the table.

What the plot inherits from the run and does not recompute. The leave-one-out sizing accuracy (RMSE and worst absolute error in bases, measured on the standard's own fragments, whose lengths are known by construction), the size-standard channel detection and its evidence, the peak resolution in bases, the peak-detection threshold, and the exact accounting of peaks found, merged, off scale, below and above the analysis range. Those are the table's readouts; the plot is drawn from the same single analysis, so a number on one is the number on the other.

Why it is a canvas rather than a rail panelCopy Anchor Link

A fragment run is 500 bases of axis in which the thing you are reading is a few bases wide: a cluster of alleles, a stutter shoulder one repeat below a real peak, a pull-up artifact showing through from a neighbouring dye. At rail width those are one smear. On the stage they are separate marks you can zoom into, which is the whole reason this surface exists separately from the table.

What to look atCopy Anchor Link

• A peak that is marked off scale has saturated the detector. Its position — and so its size — is still read normally; its height is a floor rather than a measurement.
• A region with no dashed verticals is a region where the sizing curve is interpolating over a wide gap in the standard. Sizes there are the least well supported on the plot.
• The ruler channel is labelled in the legend. Hiding it is often the first thing worth doing: it is the tallest thing on the plot and it is not your sample.