2
$\begingroup$

So, I have calculated the frequency of all SNPs (single nucleotide mutations) in some samples and have also found the nucleotides on the 3' and 5' sides of the mutated nucleotide. So I have a list of 192 possible permutations like this one:

AAA -> ACA 43
AAA -> AGA 35
AAA -> AAA 134
.
.
TGA -> TAA 345
TGA -> TCA 25
TGA -> TTA 97
.
.

What is the statistical approach to show that a particular set of 4 of these mutations is significant?

I have little knowledge in statistics so I welcome a for-dummies explanation!

$\endgroup$
  • $\begingroup$ Please do not suggest it is a good assumption to assume you are an idiot. It is wrong and you should work on knowing it better. I edited your post. $\endgroup$ – Remi.b Sep 19 '16 at 14:50
  • 1
    $\begingroup$ If you are wondering about statistical significance, it means that you want to do a statistical test. In order to make a test, you need a question to test. What is your question? Is it "Is a given SNP under selection?" or "Does the SNP really exist or is it just a sequencing error?" $\endgroup$ – Remi.b Sep 19 '16 at 14:53
  • $\begingroup$ Hi, the stupidity I was referring to was statistics-specific :) But, you are right. Anyway, I am trying to prove that the SNP is under selection. Specifically, that there is a higher frequency of TCA>TTA and TCT>TTT. $\endgroup$ – nikonan Sep 19 '16 at 19:20
  • $\begingroup$ Cool. You should clarify in your question that you are trying to detect selection. Also, what samples do you have exactly. Do you have different species or a single species? Do you have a structured population or a panmictic population? Do you have candidate genes (or SNPs)? $\endgroup$ – Remi.b Sep 19 '16 at 20:59