Question 18 of 20 Advanced Scenario
How would you decide whether confluence is an acceptable proxy for proliferation across treatments?
Short answer
Confluence measures the area cells cover, which depends on cell number and on cell size and shape. It is acceptable only if the treatments do not change morphology, do not cause detachment or multilayering, and the wells stay below the point where growth saturates. Validate it against counts for each treatment before trusting it.
Why
Live-cell imaging platforms report percentage confluence cheaply and continuously, which is why it is tempting. The assumption underneath is that area is proportional to cell number. Treatments break that assumption in predictable ways.
Cells arrested in G2 or undergoing stress often spread and flatten, so area rises while number does not. Cells that round up before division or during apoptosis cover less area at the same number. Detaching dead cells vanish from the image entirely, which makes a cytotoxic treatment look like growth arrest.
Above about 90 percent coverage the measure saturates, and any further division is invisible. Transformed lines that pile up multilayer, breaking the relationship completely. The practical decision is empirical rather than theoretical.
Run the treatments once with both confluence and an orthogonal count, and check whether the relationship between area and number holds for each arm. Where it does, confluence is a cheap continuous readout. Where it does not, combine it with a nuclear label so cells are counted rather than measured by area, and record detached cells separately.
A strong answer covers
- Confluence conflates cell number with cell size and shape
- Detachment removes dead cells from the image
- Saturation above high coverage and multilayering
- Validate against counts separately for each treatment
- Nuclear labeling counts cells instead of area
Common mistakes
- Reading reduced confluence as cell death without checking for detachment
- Using confluence in wells that reach full coverage before the endpoint
- Validating the proxy in vehicle wells only and assuming it holds under treatment
The interviewer might ask next
Which additional measurement would turn confluence imaging into a defensible proliferation assay?