Hi,
Im confused on how to present my result for western blot for the paper. I did the experiments thrice by licor and quantified and for each time the machine gave me different values for the same samples. Since the SD is too much Iam not getting a significance in the P value even though the blot and the quantification shows a drastic effect of my experiment. How should I present my results for the paper with the SD and P values included for fold difference
I have attached a sample in the xl sheet. Any help would be appreciated.
Thnx in advance
Jaya
If I'm reading your spreadsheet correctly, you have 3 experiments, in each of which you have measured triplicates. You can calculate the means for each of the experiments, then calculate the mean of the means and the standard error of the mean.
Thank you. what i had given was the fold difference, sorry I wasnt clear about it. Is it better the give the fold difference or the mean for the normalized values.
If I give the fold difference, do I have to give the SD on that.
If I give the mean, certainly, Iam going to give the SD or SE, in that case would the reviewer ask for the P value. Since the SD Is too high, Im not getting significance in some cases.
My suggestion is that you should use your raw data and block ANOVA, where the block includes the data of a single experiment. Do not forget to verify the assumptions for ANOVA; it is critical to check out whether they are met.
When calculating the fold change, which is a ratio, remember that both the numerator and denominator have experimental error associated with them. In other words, each has a mean and a standard deviation. You must therefore apply the correct calculation for the propagation of error for a ratio to calculate its standard deviation. The formula may be found here:
http://www.fas.harvard.edu/~scphys/nsta/error_propagation.pdf
This is a good question and is related to one of my pet peeves that I encounter as a statistical reviewer for medical journals. I recommend raising the standard of presenting such data and show the raw data. The common practice of showing two summary statistics to summarize just a few data points is not sound. Even if the actual sample size is larger, say 10-20, the raw data in a stacked dot plot is more informative. Although extremely common in the literature, reporting of means with bar graphs is very inefficient (some related references below). SEM is rarely justified. Further, fold changes are typically not normally distributed and reporting the mean and testing for mean differences may be inappropriate. Data transformations or nonparametric tests may be needed.
Weissgerber TL, Milic NM, Winham SJ, Garovic VD. Beyond bar and line graphs: time for a new data presentation paradigm. PLoS Biol. 2015;13(4) doi: 10.1371/journal.pbio.1002128. PMID: 25901488 PMCID: PMC4406565
Vaux DL. Basic statistics in cell biology. Annu Rev Cell Dev Biol. 2014;30:23-37. doi: 10.1146/annurev-cellbio-100913-013303. Epub 2014 Jul 2. Review.PMID: 25000992
- Badges
- Science topic