Raw Signal

The four channels as the sequencer recorded them, before analysis subtracted baselines, deconvolved the dyes and shifted the channels into register.

What it doesCopy Anchor Link

The chromatogram you normally read is not what the instrument measured. Between the detector and the trace on screen, the analysis software does at least four things: it subtracts a per-channel baseline, it applies a colour matrix that separates the four dyes' overlapping emission spectra, it shifts each channel along the time axis to correct for the different mobilities the dyes give the fragments they label, and it discards the dye front — the early, enormous, sequence-free signal from unincorporated terminators running ahead of everything else.

Each of those steps can go wrong, and when one does the result is a trace that looks bad while the underlying run was fine. This section shows the signal from before any of them happened, so you can tell the two cases apart.

When you'd use thisCopy Anchor Link

• When a read looks bad and you need to know whose fault it is. Clean raw channels under a messy analyzed trace mean the run was fine and the analysis was not — often a wrong dye set or mobility file. Messy raw channels mean the chemistry or the capillary.
• When one base is systematically over- or under-called. A channel whose baseline sits far above the others is a dye or a detector problem, and it is invisible in the analyzed trace precisely because baselining removed it.
• When the read starts much later than you expect. The number of scans analysis discarded off the front is the size of the dye front, and a large one is a run that took a long time to resolve.
• When comparing two reads from the same plate. Raw peak values are comparable between runs on one instrument in a way that analyzed, rescaled traces are not.

What this computesCopy Anchor Link

• The four raw channels, plotted as an envelope over the instrument's own scan axis. Each channel is assigned to a base by the file's own recorded base order, not by position
• Raw scans — how many samples the detector took over the whole run
• Peak raw value, the largest sample across all four channels, undownsampled
• Per-channel baselines, each the median of that channel over the whole run. The median rather than the mean because peaks are one-sided: a mean is pulled upward by the signal it is supposed to sit under
• Baseline spread, the highest channel baseline minus the lowest — one number for "is one dye channel sitting high"
• Scans discarded from the front, the difference between the raw and analyzed scan counts, which is the dye-front region analysis removed

What it refuses to invent: a file that carries no raw channels gets no plot and no substitute. The section says which format it is and that the format stores only the analyzed signal.

Where it livesCopy Anchor Link

The Raw signal section of the Chromatogram panel, reporting on the active track, below the quality profile — the two belong together, because they are the two views that read the signal rather than the base calls. Its header carries a detach control, which is worth using here: the view is meant to be read alongside the analyzed trace, not instead of it.

Inputs and outputsCopy Anchor Link

In: the active trace's raw channels, read from the file. AB1 files carry them; SCF and ZTR do not, and a read in one of those formats is not given a substitute.

Out: the four channel envelopes, the raw scan count, the peak raw value, the per-channel baselines and their spread, and the number of scans analysis discarded from the front.

ControlsCopy Anchor Link

The scienceCopy Anchor Link

Why the raw and analyzed traces are on different axes, and cannot be overlaid. Analysis mobility-shifts each channel by a different amount, because the four dyes change the electrophoretic mobility of the fragments they label by different amounts, and a fragment's dye should not change where it appears. That correction is what makes the analyzed peaks line up into a readable ladder — and it means a raw scan index and an analyzed one are not the same coordinate. Base positions therefore do not exist on this plot, and drawing them would assert a correspondence that is not there.

Why the raw trace looks nothing like the analyzed one. The dyes' emission spectra overlap heavily, so every detector channel sees some light from every dye. The colour matrix inverts that mixing. Before it is applied, a single base's peak appears in all four channels at once, with the correct one merely tallest — which is why the raw trace reads as four broadly similar humped curves rather than four sharp interleaved ladders.

Why the plot is an envelope. A run is ten to twenty thousand scans, more points than the plot is wide. Each plotted point is the maximum within its bucket rather than the mean, so a peak survives the downsampling; a mean would flatten exactly the features worth looking at. The bucket size is printed under the plot, and values read off it are peak heights.

The dye front is the unincorporated dye-labelled terminators, which are small and run ahead of every real fragment. It carries no sequence, saturates the detector, and is the first thing analysis removes.

LimitationsCopy Anchor Link

• Only AB1 files carry raw channels. SCF and ZTR store the analyzed signal alone. This is a property of those formats, not of the instrument that wrote them.
• The plot is an envelope, not the signal. For a single suspicious scan this is the wrong view; it is built to show the shape of a whole run.
• Raw values are instrument units and are not comparable across instruments or across dye sets. Comparing them between two reads from the same plate is meaningful; comparing them to a number from another lab is not.
• A clean raw trace does not mean a correct read. It means the instrument did its job. Whether the right molecule was loaded is a different question, and this view cannot answer it.