I am calculating a certain statistic(relative entropy) of a certain sequence alignment over DNA sequences. The statistic has a certain distribution in the entire genome of which I only have 1000 sequences of length 1000. I want to test the significance of a certain value of the statistic but I obviously do not know the distribution of the statistic so I can't use t-test and such and I also have a sample of size 1000 from the population. I guess I could use non-parametric methods like bootstrapping and such to approximate the p-value of a certain statistic value but the research papers that I have been reading seem to use methods from Large Deviation Statistics and Fast Fourier Transforms. I have not have time to learn these concepts on my own and just need a nudge in the right direction. Any ideas ?

  • $\begingroup$ Bootstrap sounds good to me. $\endgroup$ – user10525 Apr 18 '12 at 10:43
  • $\begingroup$ Any ideas on something faster and more precise ? $\endgroup$ – Farhad Apr 18 '12 at 23:07

Your Answer

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

Browse other questions tagged or ask your own question.