I am using qBase method (Hellemans et al Genome Biology 2007 8:R19) to analyze real time PCR data of untreated cell cultures and in response to hormone treatment. After getting the fold change (normalized relative quantification, which means, the 2^dCt of the target gene divided by the 2^dCt of references genes), it is intuitive for me to show data by plotting the log2(fold change), which is clearer to see up or down regulations. But first, I scaled all samples to the control FC value, by calculating the mean of all replicates for each sample (biological replicates) and divide all means of each sample by the mean of the control sample (untreated), so the control sample fold change is = 1, and treated samples will have >1 values for up- and
It sounds like you are trying to determine standard error for your technical replicates. If you are interested in the errors due to technical variation, you would then have to treat each technical replicate as a separate data point and calculate fold change for each data point to then do statistics. This, of course, won't give you biological significance, and can only tell you how good your technique and instruments are.
However, since you have biological replicates as well as technical replicates, I'd recommend to focus instead on performing statistics on your biological replicates. Based on what you've done you have the means for each biological replicate, and by grouping in "treated" and "untreated" groups, I would imagine you would still have enough data point to calculate standard error of the mean or conduct a t-test if needed.
When I did gene expression studies with fish lymphocytes, I found the Relative Expression Software Tool (REST) very helpful. It provides an accepted itterative comparisons of all replicates and includes also some plotting options. It's free and can be downloaded at:
http://www.gene-quantification.com/rest.html
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis