3D-proteine structure: Homology modeling versus X-ray chrystallography

[C. Rebeiz]

>> . . . I wonder
>> however, about the relationship of a laboratory-grown crystal of a
>> membrane-bound solubilized protein to the native protein in situ. 
>> Constantin A. Rebeiz
>> 240 A, PABL
>> 1201 W. Gregory
>> Univ. Illinois, Urbana, IL 61801
>> 
>True, many of us do -- but at least we know the real fold of that 
>polypeptide under at least one set of conditions.
>
>It is often not mentioned that the "folding problem" is actually the
>prediction of the structure of a particular sequence in a particular
>                                                     ^^^^^^^^^^^^^^^
>environment.  If this were not so, proteins would not denature, and
>^^^^^^^^^^^
>we know that some of them are very sensitive to things like pH,
>temperature, salt, . . . 
>
>This makes the "folding problem" something more than an exercise in
>pattern detection and matching.
>
>Lynn Ten Eyck
>teneyck at sdsc.edu
>
In that case the debate that has been going in this forum, on on the
relative merits of X-ray data versus 3D-modeling is trivial since neither of
the two techniques can be related with absolute certainty to the state of a
protein in situ.  I must add however that twith this uncertainty in mind,
dynamic 3D-modeling (amber) and docking techniques for example, do permit
the  derivation of additional 3D-models derived from a homology model,
which may be more compatible with known experimental data.  In other words
the dynamic choices offered by 3D-modeling which are not offered by the more
rigid X-ray data, should not be iognored or minimized. 
>
Constantin A. Rebeiz
240 A, PABL
1201 W. Gregory
Univ. Illinois, Urbana, IL 61801