Hi everyone,
What statistical test would be most appropriate for this experiment?
I want to compare two conditions against one another at different time points, e.g. mRNA1 vs mRNA2 at 24 hours, mRNA1 vs mRNA2 at 48 hours, etc.
Two conditions: standard eGFP mRNA (n=3 separate IVT syntheses) and modified eGFP mRNA (n=3 separate IVT syntheses) were transfected into HEK cells and protein levels were detected by flow cytometry at 24, 48, 72, 96 and 168 hours.
I did flow cytometry analysis and detected eGFP intensity median at each timepoint for 3 samples in each of the two conditions (median expression of protein for modified mRNA vs standard mRNA, and analysed in Graphpad prism.
The analysis was with following parameters:
Test name: Unpaired t test with Welch correction
Variance assumption: Individual variance for each group
Multiple comparisons: False Discovery Rate (FDR)
Method: Two-stage step-up (Benjamini, Krieger, and Yekutieli)
Desired FDR (Q): 1.00%
Number of tests performed: 5
Number of rows omitted: 0
Is this an appropriate test?
Thank you,
Maria
Hi Mary,
although I am not a biomedical researcher and am not familiar with the practices in this field, your problem description tells me that you should use a mixed ANOVA to avoid alpha error accumulation. Alpha error accumulation can occur if you repeatedly use t-tests on the same population.
Mixed ANOVA is most commonly used in research where two or more groups are to be compared across multiple time points (e.g., a drug group and a placebo group tested multiple times during the trial period). It "automatically" analyzes within-group differences (over time) and between-group differences, which is why it is also called a between-within ANOVA.
I hope I was able to help you and wish you much success with your research project!
Matthias
Hi @Matthias, my supervisor suggested that two way anova between every triplicate in every timepoint may lead to Geaphpad prism analysing this data in such way that the data gets normalised to average and therefore, not showing much statistical difference. I did two way anova initially, but had to change to T-test, but I don't think it is the right type of test.
It depends on your problem assumptions since there are two main types of tests: parametric and non-parametric. When compared to non-parametric testing, parametric tests make specific data assumptions while non-parametric tests do not.
- Badges
- Science topic
- Similar topics
- Phytochemistry
- Extracts