Question 8 of 9 Intermediate Troubleshooting
Replicates vary widely only at low template concentration. How would you investigate?
Short answer
At low copy number, stochastic sampling itself causes variation. Investigate whether the spread matches sampling statistics before diagnosing pipetting or instrument failure.
Why
If template molecules enter wells approximately randomly, the count is close to Poisson: a mean of one copy leaves about e^-1, or 37%, of wells with zero molecules. Even excellent chemistry cannot make all near-LOD replicates positive. Reverse transcription and extraction add additional variation before PCR.
How to work it through
- Plot Cq and non-detection frequency by input level.
- Inspect whether replicates were independent extractions, RT reactions or only PCR wells.
- Run a controlled dilution series with enough replicates near the detection limit, evaluate low-copy reference material, contamination and plate position, and check efficiency away from the low-copy boundary.
- Report detection probability and the distinction between LOD and reliable quantification.
An example
At about one target molecule per reaction, some wells remain negative and positive-well Cq values spread. If abundant controls remain tight, that can be expected sampling behavior, not evidence that those negative wells were technically mishandled.
Where the answer stops holding
Clumped molecules violate ideal Poisson assumptions. Inhibitor gradients and adsorption to plastics can make variation worse. Do not replace missing Cq values with a fixed arbitrary cycle and then compute a simple mean.
A strong answer covers
- At low copy number, stochastic sampling itself causes variation
- Investigate whether the spread matches sampling statistics before diagnosing pipetting or instrument failure
- Plot Cq and non-detection frequency by input level
- Inspect whether replicates were independent extractions, RT reactions or only PCR wells
Common mistakes
- Treating technical replicate counts as biological sample size
- Excluding late or negative wells without a predefined rule
- Claiming precise fold changes below LOQ
The interviewer might ask next
How would you report a target detected in four of twelve technical wells?