There are numerous ways to compare 3D structures. Consider alpha-carbon
rms differences upon superimposition. Consider sequence similarity
based on many different kinds of criteria (identity, size, chemical
similarity, average burial, etc). If you are trying to classify a
protein's family, consider the latter. If you are modelling, consider
the former, as well as the similarity between core residues (assuming
you are working with a globular protein).
Also use motif searches to guess at functional similarity. For a
quick guess as to whether or not they are structurally analogous,
perform a secondary structure alignment, much like a pairwise sequence
alignment, looking for >70% identity (use internet resources for
secondary structure determination, then align by hand). Again, it's
somewhat difficult to answer without knowing what you are trying to
accomplish. Thanks for your question.
Wright State University
Philippe CHAVATTE wrote:
>> What is the best methodology to compare two protein 3D structures (pdb
> format) ?
> What criteria permit to affirm their similarity ?
> Many thanks for your help.
>> Philippe CHAVATTE
> Institut de Chimie Pharmaceutique Albert Lespagnol
> BP 83
> 59006 LILLE CEDEX
>> E-Mail : pchavatt at phare.univ-lille2.fr