I have a question about which statistical test to use for a particular research question. I am studying protein binding motifs (amino acid sequences that other proteins bind to), typically these motifs are at the end (C terminus) of proteins. I want to know if specific amino acids are significantly enriched at any of the last six positions (C terminus) for all proteins in a given organism. If this is the case, then this would support the hypothesis that various binding motifs are widespread throughout the proteome.
For example, let's say that I want to look at the last position (C terminal position) of all proteins and ask if there are any amino acids that are more abundant than in the rest of the proteome. First, the frequency of all 20 amino acids at that final position only would be calculated (using all proteins), these are our observed frequencies. Second, the frequency of all 20 amino acids at all positions would be calculated (again using all proteins), these are our expected frequencies. Then for each amino acid, I want to test if the observed frequency is significantly different than the expected frequency.
I believe that I could use a Chi-Squared goodness of fit test, but I am tentative to use this test as I am not using a simple random sample of sequences, but rather I am just restricting the position from which the frequencies are calculated.
Are there any suggestions as to which test I should use? Thanks for any suggestions ahead of time.
