Designing RNAi Experiments
1. Maximize the signal-to-noise ratio in your phenotypic assay
The causes of noise or nonspecific signals in assays used to evaluate the effects of RNAi are as diverse as the assays themselves. Some of the most common variables that have a strong influence on the signal to noise ratio include:
- the number of cells evaluated
- the choice of reagents such as antibodies and/or substrate
- the timing of the assay after siRNA delivery or addition of substrate
- the turnover rate of the targeted gene product
- the effectiveness of the siRNA
2. Identify and minimize edge effects of culture plates
The position of samples within multiwell culture plates, especially 96 and 384 well plates, can have a surprisingly strong influence on assay results. This may be due to a combination of effects on cell growth and the assay itself. [Malo N, Hanley JA, Cerquozzi S, Pelletier J, Nadon R Nature Biotechnology 24 (2): 167-175] Typical of cellular assays on plates, Ambion has seen significantly more variability in cell growth and/or assay performance in outer ring and corner wells. The degree of variability was not necessarily consistent from plate to plate whether evaluated on the same day or on different days. This variability can be managed by improving plating technique and by statistical methods to normalize the results. Listed below are suggestions for eliminating or mitigating plate position effects on RNAi data:
- If your throughput requirements allow for fewer samples per plate, consider avoiding all edge wells in multiwell plates. Instead load those wells with culture medium only or with nontransfected (or non-electroporated) cells.
- Position replicates in the plate so that data from them can be used to evaluate position effects. For example, divide the plate into quadrants and include replicate samples or, at a minimum, positive and negative controls, in different quadrants of a single plate. Alternatively, include replicates in different locations across several multiwell plates.
- Use mathematical or statistical methods to minimize edge effects. Strategic placement of well-defined positive and negative controls can potentially be used to spot faulty data or to normalize data based on location within or among plates. Setting separate requirements for retesting or evaluating ring well data might also be implemented.
3. Establish standard operating procedures for handling and growing cells
When establishing cell handling and culture procedures, consider even obscure details such as the culture plates used, lot of culture medium, timing of trypsinization for detachment, and placement within the incubator. Variability in the more obvious parameters such as the passage number of the cells, their confluence when harvested for experiments, and the method used to keep the cells suspended as they are distributed should, of course, be minimized.
3. Establish standard operating procedures for handling and growing cells
When establishing cell handling and culture procedures, consider even obscure details such as the culture plates used, lot of culture medium, timing of trypsinization for detachment, and placement within the incubator. Variability in the more obvious parameters such as the passage number of the cells, their confluence when harvested for experiments, and the method used to keep the cells suspended as they are distributed should, of course, be minimized.
4. Establish standard operating procedures for assay conditions
Similarly, the procedures for conducting the assay should be standardized. This includes all reagent vendors, reagent volumes, the timing and conditions for siRNA delivery, and assay performance. Where possible, the same lot of reagents—especially serum containing reagents—should be used. Method of pipetting, and the calibration status of pipettors, liquid handling systems, and data collection equipment, can all affect variability if not standardized.
5. Monitor variability and assay performance using carefully chosen controls
Even with an optimized assay performed using standard procedures at each step, continue to monitor your experimental results for higher than expected variability by including positive and negative controls for each step of the assay. Finally, consider retesting samples if necessary.
