percentage of amino acids, normalize by background amino acid probability

Question

[bioinfo8]

Hello, I have 11 proteins in two datasets: with 5 in 1st and 6 in 2nd dataset. I have calculated the amino acid percentage for three amino acids (A, R, N) in them as shown in the attached file from excel

I want to normalize them by background amino acid probability of all proteins in two datasets and then conclude, for e.g., basic amino acids (R) are enriched or hydrophobic amino acids (A) enriched. Kindly help.

Thank you!

Answer 1

[abascalfederico]

Percentages and frequencies are already a form of normalization, aren't they?

May be what you mean is how to test whether proportions of certain types of aa are significantly different in the two sets . A Fisher's exact test will do the job.

[bioinfo8]

Hello, thank you for the answer. Can you please guide me further and refer some related papers. Thank you.

[abascalfederico]

Papers related to what? Fisher's exact test?

[bioinfo8]

This kind of analysis using the test, thanks. I think I can do with R.

[abascalfederico]

I am not aware of references doing this kind of analyses, sorry. But running a Fisher's exact test to compare proportions of amino acids is straight forward. Yes, do it in R.

[bioinfo8]

Ok, I will try to find out some papers and consult you for further analysis.

[bioinfo8]

These are few papers: 1) Paper 1 2) Paper 2 Now I am stuck on how to arrange data so that R can read it, then I can use fisher.test function and wish to get output in the mentioned papers. Kindly guide. Thanks.

[abascalfederico]

I will suggest to start with some basic R tutorial, there are lots in internet. Once you are familiar, prepare a matrix of 2x2 for your proportions and run the Fisher test.