Every QAID microarray slide carries assay controls alongside your sample, and the QAID App reads them before it reports a single species. It is worth understanding why, because the controls answer a question that comes before “what is in this sample?” — namely, “did this run work?”
The question underneath the result
A negative result has two possible explanations. Either the target was not in the sample, or the test did not work. Those look identical on the slide: in both cases, the probes for that species stay dark.
That ambiguity is the oldest problem in analytical testing, and controls are the standard answer to it. A control is something you already know the answer to, carried through the same process as your sample. If the control behaves as expected, the process worked, and a dark probe means what you hoped it meant. If the control does not behave, nothing else on the slide can be interpreted.
This is why the QAID App evaluates controls first. A run whose controls fail is flagged, not reported as a clean result. A technical failure should never be allowed to look like a passing sample.
What the controls are watching
Controls on a microarray are not a single check. Between them, they cover the stages where a run can quietly go wrong.
- Did usable DNA make it out of the sample? If extraction recovered little or nothing, from a heavily processed material for example, the run cannot answer anything about species. A sample producing no signal anywhere is telling you about your extraction, not about your supplier.
- Did the chemistry run? Preparation and labelling have to work for target DNA to be detectable on the slide. Controls show whether that part of the process behaved.
- Did the slide and the scan behave? Alignment, focus and scanner settings all affect what is read at each probe position. Controls placed on the slide show whether the scan is being read correctly.
- Is the background where it should be? Every slide has a background level, and a call is made by comparing probe signal against it. An unusually high background is a warning that faint signals cannot be separated from noise on that slide.
- Was anything carried over? Signal where there should be none is as informative as signal where there should be some, and points at contamination rather than at the sample.
Pass, flag, repeat
In practice the QAID App resolves a run into one of three situations.
Controls pass. The run behaved as expected, and the species results are reported with their confidence metrics. You are reading a result about the sample.
Controls fail. The App flags the run instead of presenting species results. The right response is to repeat, usually from the extraction step, rather than to read the flagged output as a negative. Nothing about the sample has been established.
Controls pass, but the run is marginal. A high background, or weaker-than-usual control signal, can leave strong calls trustworthy and faint ones uncertain. This is the case worth training a QA team on, because it is where judgement is needed: a clear identification on a marginal slide usually stands, while the absence of a faint target on the same slide is not a conclusion.
Controls and confidence are not the same thing
These two are easy to run together, and they answer different questions.
The assay controls describe the run. They tell you whether the slide as a whole can be interpreted. They apply to every result on that slide equally.
The confidence metrics describe one call. They summarise how strongly and how consistently the probes for that target responded, relative to the background measured on that slide. Because several probes stand behind each target, a call supported by agreement across probes reads differently from one resting on a single strong spot.
You need both. Strong confidence on a failed run means nothing, because the run cannot be interpreted. Passing controls with a borderline call means the run was fine and that particular call deserves a second look.
Reading a result in the right order
A habit worth building in your QA procedure:
- Check the control outcome first. If the run is flagged, stop and repeat. Do not read the species panel.
- Then read what was identified, with the confidence metric for each call.
- Then read what was screened for and not detected. On a run with passing controls this is real information; on a marginal run, treat faint absences with care.
- Then decide. Release, hold, retest, or send for confirmatory testing. The App supplies the evidence; the decision is your team’s.
Written that way, the control outcome stops being a technical detail buried in a report and becomes the gate it is meant to be.
Why this matters outside the lab
When a customer or an auditor questions a result, the useful answer is rarely louder insistence on the conclusion. It is the ability to show how the result was produced: that the run’s controls passed, what the confidence behind each call was, and what the supporting signal data looked like. Because those sit with the result in the QAID App, that conversation is a matter of opening the record.
It also works the other way. A laboratory that repeats flagged runs instead of reporting them is a laboratory whose negatives mean something. That is the quiet reason controls earn their space on every slide.
More on how a scan becomes a result in inside the QAID App, and on how products are validated in validation & quality.