I'm measuring genetic diversity from every individual in a sample using two different types of genetic information. These types of genetic information have different properties, and mutate independently of one another, yet will always be inherited together when an individual reproduces. Should I consider these different genetic measures as independent because they change independently? Or should I consider them dependent because they come from the same individual?


1 Answer 1


You should divide it into two parts.

Use the data and test for independence, ignoring the source to the data. Use the domain knowledge in order to think about reasons to the observed dependence or independence. The classical way to test that is the Chi-Square test of independence.

Both justifications you mentioned sound possible. Let the data decide which one is true.

You should note that if your data is small or biased, it may not represent the real relation. If you know of different properties in the data (e.e.g, other genetic information) that must be independent/independent, check them in your data set as a sanity check for being not representative.

  • $\begingroup$ Thanks for the answer. It's better than I expected because it justifies my uncertainty, rather than it being just one situation or the other. $\endgroup$
    – anjama
    Dec 22, 2017 at 2:13
  • $\begingroup$ It is hard to be certain a-priori when dealing with complex data. However, testing for independence in the data set will tell you where you stand. $\endgroup$
    – DaL
    Dec 22, 2017 at 6:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.