After doing a selective neutrality test of the molecular markers I worked with, it turned out that 30% of the markers are subject to selection and the remaining 70% are neutral, should I do my study with those who are neutral only or should i use all of them?
It is better to work with neutral and non-correlated SSR markers. But, you cannot say for sure that chosen by you SSR loci have such properties before you complete your study.
As a clarification: I am assuming that you mean 30% of your SSRs had statistically significant evidence of departures from neutrality, correct? Unless you have a genome and know they are from known region, you don't "know" they are under selection, just that the statistical test indicates they might be. Thus, immediately removing them from the dataset would be an over-reaction. Those tests are prone to Type I errors.
It depends on the analysis you want to perform. Some measures of genetic differentiation (e.g Fst, Bayesian clustering) assume markers are neutral. Others (such as a PCA) do not. Some are robust to departures of neutrality: check the assumptions first.
Depending on how many markers you have, run the analyses with the full suite, just those "under selection", and those presumably neutral: see if you get a different answer.
Hope this helps.
thank you for all your answers, I tried to work with and without non-neutral markers, I didn't find any significant difference between the two cases, so in the end I kept all the loci for analysis.
- Badges
- Science topic