Back to protocols page
Here are some basic statistical considerations for the AmpliCot assay:
- Standards of known diversity will help you translate measured Cot values into absolute diversity measurements. Traditional Cot values were interpreted using extrapolations from very simple standards. It is possible to use plasmid clones of the sequences you are studying (e.g. T cell receptor genes) to make simple standards of known diversity. It is more difficult to make reproducible complex standards of known diversity, but this would be useful because interpolated measurements are more reliable than extrapolated ones. One possiblity for T cell receptor measurements is to use known numbers of diverse cells (e.g. pooled naive cells) to make more complex standards for the assay.
- To be useful, diversity measurements require that the sample you are studying contains the full diversity of the population you are interested in. In the case of lymphocytes, this may require larger samples than you are used to working with. No matter what sequences are being measured, it is useful to titrate the sample size to be sure that you are working at a plateau, where further increases in sample size do not increase the measured diversity.
- A final basic consideration is measurement error. If standards are included in the annealing reaction, then day-to-day differences in salt or dye concentration in the buffer are less likely to contribute to error. The major source of error, in the case of RNA-based assays, is going to be RNA extraction, which is difficult to control. It is worth processing identical aliquots of the same sample on different days to measure your error due to this step (inter-assay variability). One way to reduce the magnitude of this variability is to make sure that your sample sizes are large enough that you are operating far out on the plateau (see the paragraph above). In this way, you will be buffered against small sample-to-sample variations in efficiency of the RNA extraction, RT and PCR steps of the protocol.
We hope to publish a more sophisticated statistical analysis of AmpliCot in the future.
Last updated September 5, 2006