Question 9 of 20 Fundamentals Interpretation
What does the Z prime factor tell you about a screening assay?
Short answer
Z prime combines the separation between the positive and negative control means with the variability of both, giving one number for how well a single-point screen can tell a hit from background. Values above about 0.5 are usually treated as acceptable, and it says nothing about potency measurements.
Why
The statistic subtracts three standard deviations of each control from the gap between their means, then divides by that gap. A large gap with tight controls approaches 1. Overlapping controls give a negative value.
Iversen and colleagues compared signal window, Z prime factor and assay variability ratio and recommended Z prime as the preferred measure of performance for screening assays. The same paper makes the limitation explicit: none of these measures is adequate for characterizing concentration-response assays. That is the point candidates most often miss.
A screen separates active from inactive at one concentration. Ranking compounds by potency is a different task, and Eastwood and colleagues introduced the minimum significant ratio for it, the smallest fold difference in potency between two compounds in the same experiment that can be treated as real. Z prime also depends only on controls, so it cannot see systematic plate effects, edge effects or drift across a long run.
Those need plate uniformity studies and control wells spread across the plate.
A strong answer covers
- Combines control separation with control variability
- Above about 0.5 is commonly accepted
- Recommended over signal window and variability ratio for screens
- Does not characterize concentration-response assays
- Minimum significant ratio is the metric for potency comparisons
Common mistakes
- Using Z prime to judge whether two IC50 values differ
- Reporting Z prime from one plate and assuming the whole run is fine
- Improving Z prime by picking extreme controls that no real sample resembles
The interviewer might ask next
How would you decide whether a twofold potency difference between two compounds is real?