Ya,       

     your question is not properly posed. You must specify more details about your case in order to get a proper answer. For the 'mutant sites' (this definition by itself is not very clear - you mean sites that accept mutations in a protein family?) you likey talk about a protein bearing mutations relative to other known proteins of the same family. In this case, multiple sequence aligments will uncover the most frequent mutation sites (depending on how many sequences you can align). The same kind of answer holds for the question about uncovering residues in an enzyme active site. If you can recognize sequence relationships to homologs it will be very easy to transfer such information from these to your enzyme. In case you have just a 'new' sequence unrelated to anything else in the sequence/structure databases, it's going to be tricky. Likely, you will need to integrate your sequence information with some enzymology work in vitro.

Martino Bologensi

Ya,       

     your question is not properly posed. You must specify more details about your case in order to get a proper answer. For the 'mutant sites' (this definition by itself is not very clear - you mean sites that accept mutations in a protein family?) you likey talk about a protein bearing mutations relative to other known proteins of the same family. In this case, multiple sequence aligments will uncover the most frequent mutation sites (depending on how many sequences you can align). The same kind of answer holds for the question about uncovering residues in an enzyme active site. If you can recognize sequence relationships to homologs it will be very easy to transfer such information from these to your enzyme. In case you have just a 'new' sequence unrelated to anything else in the sequence/structure databases, it's going to be tricky. Likely, you will need to integrate your sequence information with some enzymology work in vitro.

Martino Bologensi

Hi, are you looking for a novel enzymatic site present in a known protein? The active sites of many enzymes have already been characterized and you can do sequence comparisons with those known enzymes.

The usual approach to this kind of problem is to look for homology with other similar proteins. If one (or some) of the homologues have been characterised, your starting point is to look for the equivalent residues in your unknown protein that correspond to functional parts of the known protein.

If your homologues  are also of unknown function, the problem is harder. As a first approximation, look for regions that show higher conservation as these are likely to include substrate and regulatory binding sites. Actives site residues themselves tend to be conserved, but what is close spatially may not be very close in the sequence.

When you talk about "possible mutation sites" I'm not sure what you mean. Are these sites that you could mutate to alter the function of the protein, or sites that vary within a protein and don't alter function appreciably?