NeuroQP now supports cell counting for projects that do not have a usable DAPI staining.
That matters for a very practical reason: a lot of labs already have data. Sometimes the experiment was designed without DAPI. Sometimes the DAPI channel is missing. Sometimes the nuclear stain exists, but is not good enough to use as a reliable cell-detection source. The dataset is real, the biology may be useful, and nobody is going to restain the past.
If you are planning a new experiment, our default advice is still simple: if you can include a good DAPI or nuclear stain, do it. DAPI-based workflows are generally more complete because they give NeuroQP a visible all-cell reference population.
But if you do not have DAPI, the answer is no longer: sorry, come back with a different experiment.
Marker-based cell counting without DAPI can work. It just answers a narrower question. Instead of asking what percentage of all detected cells express a marker, NeuroQP can detect marker-positive candidates directly on the marker staining, classify which detections are real cells, and quantify those accepted marker-positive cells across regions.
That is the distinction this article is about.
DAPI Gives You The Denominator
In a DAPI-based workflow, NeuroQP first detects cells on the DAPI or nuclear channel. That detected population becomes the shared reference for the rest of the analysis.
After that, marker channels can be classified against those same cell positions. A detected DAPI cell can be identified as marker-positive or marker-negative. Either way, it is still part of the same measured cell population.
That is why DAPI-based workflows are so useful for brain-slice quantification. They support richer metrics:
- total detected cells
- marker-positive cells
- marker-negative cells
- percentage of cells expressing a marker
- all-cell density
- marker-positive cell density
- co-expression using shared cell identities
This is also why we have written separately about automated DAPI cell counting for mouse brain slices and cell detection for DAPI brain slices. DAPI is not just another color channel. It is often the anchor that makes the rest of the measurement cleaner.
If you want to report that 18 percent of cells in a region express cFOS, you need to know what the other 82 percent are. DAPI gives NeuroQP a practical way to define that population.
Why DAPI-Based Workflows Are More Complete
The power of the DAPI workflow is not that DAPI is glamorous. It is not. DAPI is very good at being blue and useful.
The power is that every marker can refer back to the same detected cell population.
That makes positive and negative marker identifications meaningful in the usual biological sense. A positive identification means the detected cell expresses the marker. A negative identification means the detected cell does not show enough marker signal to be called positive.
Those negative identifications are still real detected cells. They are not detection failures. They are part of the denominator.
The same shared population also makes co-expression easier to reason about. If marker A and marker B are both classified on the same DAPI-derived cell IDs, then asking whether one cell expresses both markers is straightforward. The cell identity exists before the marker question is asked.
So if the data supports this workflow, NeuroQP will generally get more complete measurements from it.
When The DAPI Channel Is Missing
Real datasets do not always arrive in the preferred form. Very inconsiderate, but there we are.
Some projects do not include DAPI at all. Some have a nuclear stain, but it is not usable enough for automated detection. Some historical datasets were collected around marker-only imaging. Some experiments were designed before anyone expected the analysis workflow to need a denominator later.
That does not make the data useless. It changes what NeuroQP can measure from it.
A marker-only image can show cells that express that marker. It usually cannot show cells that do not express it. The negative population is not quietly waiting in the background for software to discover it.
So NeuroQP uses a different workflow for these projects.
How NeuroQP Handles Marker-Based Counting
For non-DAPI projects, NeuroQP can detect candidate marker-positive cells directly on the marker staining.
Instead of detecting all nuclei on DAPI and then checking marker expression at those nuclei, NeuroQP runs detection on the marker image itself. Each marker gets its own marker-specific detection run.
That word candidate matters.
A marker channel is not always a clean list of cells. It can contain speckles, axonal fragments, bright debris, background signal, merged blobs, and other non-cell structures that look tempting to a detector and unconvincing to a human reviewer.
So NeuroQP does not treat every raw marker detection as final biology. The workflow is:
- Detect candidate marker-positive objects on the marker staining.
- Classify those candidates as real marker-positive cells or noise.
- Use the accepted marker-positive cells for downstream results and statistics.
This is not meant to be a better replacement for DAPI-based counting. It is the practical path for data where DAPI is missing or unusable.
Classification Cleans Up Candidate Detections
The biggest semantic difference is that negative classifications mean different things depending on the workflow.
In a DAPI-based workflow, a negative marker classification can mean a real detected cell that does not express the marker.
In a marker-based workflow, a rejected candidate usually means NeuroQP found something in the marker image and the classifier decided it was not a valid cell. That rejected object is not a biological marker-negative cell. It is noise, artifact, ambiguous signal, or a detection that did not pass review.
That distinction is important, but the good news is practical: NeuroQP can use the same classification approach in both settings.
For common activity-marker workflows, NeuroQP includes ready-to-use classifiers for cFos- and TRAP-type stainings. These were trained on a broad range of cFos and TRAP data, so they can often cleanly separate real marker-positive cells from speckles, fragments, and background signal in marker-based projects.
If your staining or image style is different, you can train your own classifier in NeuroQP. The workflow is the same idea: review examples, label what should count as a real cell versus noise, and let the classifier apply that decision consistently across the dataset.
That is the useful part. The software does not ask every marker-only project to fit one universal definition of a cell. It lets the classification step adapt to the data you actually have.
What NeuroQP Can Measure Without DAPI
Marker-based counting can still support useful results.
NeuroQP can count accepted marker-positive cells. It can calculate marker-positive cells per mm2. It can aggregate those counts by brain region, animal, group, or condition. If the biological question is about the density or distribution of marker-positive cells, this can be exactly the measurement you need.
This is especially useful when the target cells are visible through the marker channel itself. If the stain reveals the population of interest directly, then direct marker detection is a sensible fallback when DAPI is not available.
The valid outputs are focused:
- accepted marker-positive cell counts
- marker-positive cell density
- region-level marker-positive summaries
- group comparisons based on marker-positive cells
That is a useful measurement. It is just not the same as total cell quantification.
What NeuroQP Does Not Report Without A Cell Source
Without DAPI, or another usable all-cell source, NeuroQP cannot compute the percentage of all cells that express a marker.
The reason is simple: cells with no marker signal are invisible to the marker-only workflow. NeuroQP can quantify what the image shows. It should not invent the denominator the image does not contain.
So in non-DAPI projects, NeuroQP focuses on marker-positive counts and densities. Metrics that require an all-cell reference population are not treated as normal outputs for that workflow.
The spreadsheet might be happy to accept a percentage. The image still has to earn it.
How NeuroQP Handles Co-Expression
Co-expression also changes when DAPI is missing.
In a DAPI-based workflow, NeuroQP can classify marker A and marker B on the same DAPI-derived cells. That means co-expression can use shared cell IDs.
In a marker-based workflow, each marker has its own detection run. Marker A has its own candidate cells. Marker B has its own candidate cells. Their IDs are independent.
NeuroQP handles this with spatial overlap. If a detected marker A cell and a detected marker B cell overlap at the same position, NeuroQP can treat that overlap as evidence that they represent the same biological cell expressing both markers.
That is a different model from DAPI-derived shared cell identity, but it fits the data NeuroQP has in a marker-only project.
Use The Workflow The Data Supports
For new experiments, DAPI is still the safer default if you want the richest cell-counting workflow. It gives NeuroQP a visible cell source, a denominator, and cleaner support for percentages and shared-cell co-expression.
For existing datasets without DAPI, NeuroQP now has a useful alternative. It can detect marker-positive candidates directly on the marker staining, classify real cells versus noise, and report marker-positive counts, densities, regional summaries, and overlap-based co-expression.
That is why we added non-DAPI based cell counting.
Not because DAPI stopped being useful. It very much did not.
We added it because real labs have real data, and not all of that data comes with the stain we would have asked for if we had been standing next to the microscope at acquisition time.