to begin with, 3 proteins is not a big enough sample to say that something is conserved for functional reasons. it depends on how distant they are i suppose.

you could make some educated guesses based on the properties of the amino acid as well as where they are in the protein. furthermore, if you model the 3d structure of the protein you might find that they are important for a loop or something else structural.

I also suggest you have a look at the underlying nucleotides, to see how divergent they are.

Thank you aswad87

the 3 proteins belongs to the same family. I have done the CULSTALw alginment for each protein in a evolutionary way. So I have 3 CLUSTALw with 20 sequences approximately from amphibians or birds & reptiles (depend the protein) to primates. I have been checking the patterns and profiles got it for the 3 human proteins in each CLUSTALw to see how conserved are in the evolution.

After that I still have some regions very conserved.

I will check the predicted structure of each protein to see where these regions are located.

This is similar/related to the questions asked and answers given in these posts: A: locating motif within a protein sequence, Predicting protein-protein binding, Identifying functional regions of proteins

To give a direct answer to this question - coming from a group that works on understanding/predicting the function of short functional regions of proteins, I'd point you in the direction of the resources developed in our group, ELM, phosophELM, and the alignment-focused Conservation Scorer, and check out also the links there