Fellow Researchers,
I am currently running an experiment where primary cell lines derived from 4 patients are being cultured in 2 different conditions. I am examining expression levels of several genes in 3 technical replicates in both conditions and trying to find out whether change in culturing method has a signifcant impact on them considering all cell lines.
I am using Nested t-test for my statistical analyses, but having doubts whether or not it is the right choice since data from the same cell line in diffrent conditions is not paired. Would you suggest me to change the test or keep using it? What alternative would you suggest?
You should use an omnibus test like factorial ANOVA. This is a 2 x 4 design, so there are 8 conditions.
Multiple t-tests are not the best choice because of inflated Type I error rate. To make all pairwise comparisons with t-tests would be (8! / (2! * 6!)) = 28.
Technical replicates don't tell you anything about biological variation. You didn't say how many biological replicates you have. If only one, then you have N = 1 per condition.
If your cells are easy to grow I would do at least 10-12 biological replicates per treatment then do the ANOVA calculations and see what you get. Good luck with your experiment.
nested T-test or wilcoxon test are really good options for your experimental design i suggest you to do it in a not parametric way, until you know for sure the normality of your samples
Nested t-test is for comparing only two groups (two treatments). It is not appropriate to perform several nested t tests, comparing two treatment groups at a time. GraphPad has an excellent text/tutorial on the issue:
http://www.graphpad.com/guides/prism/8/statistics/stat_analysis-checklist-nested-t-te.htm
Realize that technical replicates are not replicates in a statistical sense. If you treat them as independent values, you violate the assumptions of the test you are trying to use (t-test) and your p-value will be too small. If you average them, you have only one value per treatment. So regardless of the statistical analysis, i would concentrate on getting more data.
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Cell Biology
- Culture Cells
- Cell Line