Dataset Of Sequences To Be Mapped On Structure

Question

[Hari]

Hi all, I would like to map the conservation of amino acids onto a structure,so we have a set of protein domains belonging to a family extracted from HMMer search.We would like to map it on a structure to see how many residues are conserved and how many are variable,now the question is should i consider isoforms and mutants in my dataset ? we have removed very redundant sequences but have decided to retain isoforms that differ even by one amino acid? is this fine? If we choose to remove them,then we lose a sequence that could potentially belong to the family? Please give me your suggestions

Thanks in advance