I am trying to determine if there is any correlation between the genome size of 10 of my species and the number of an X gene that I have identified but the sample number of 10 seems quite low for an accurate statistical analysis and i am trying to run pearson's correlation.
Hi Ahmed,
It is difficult to say it. If the correlation is strong enough, you can preserve a power of 80% with a sample size of 10. Otherwise, the analysis will lack power to accept Ha, and in general you will have a non-significant correlation.
Good luck!
Hello Ahmed,
If your resources only allow for a sample of 10, then you simply report the observed correlation (and, ideally, a confidence interval estimate of r).
Eliana Ibrahimi is correct in noting that such a sample size will not have much statistical power unless the true correlation in the population is pretty large. In fact, with an N of 10, your observed Pearson correlation would have to be at least | .64| to be significant at the .05 level (non-directional).
Good luck with your work.
Both David Morse and Eliana Ibrahimi give excellent answers. However, you don't give your research question. Is size (but in what units) size in # of genes or copy number or what? I suggest you look at other papers on Genome wide Association studies to see what you can do. I have attached a couple that might get you started and a paper of ours that contained a GWAS reference that was quite useful to us. Best wishes, David Booth
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis