Barani (barani at mace.cc.purdue.edu) wrote:
: >experimental methods, that is implicit in the system. However it can
: >provide information several orders of magnitude faster than crystallography.
: ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
: Small Correction here:- I guess the sentence should probably read as
:: "provide WRONG information several orders of magnitude faster than
: crystallography"
:: :-)
:
You can get wrong information with any technique, don't be malicious :-).
But it might take you a year just to figure out that something folds
like a B-barrel, while it might take me just half an hour. Of course
you'll be able to go into more detail than I. But by the time you
have decided that residue 155 should be mutated I have had a whole
year to mutate residues 150 to 160, thanks to my low resoulution data.
: Let me remind you a familiar Story:-
: What do 5 protein modellers make by combining
: (1) 4 pillars (2) a cylinder (3) a snake (4) two handfans (5) a barrel?
: An ELEPHANT !!!
And what do crystallographers do when you give them 50 mg of pure protein ?.
Phone in a month asking for another 50 mgs !
: On the serious side:-
: For your kind information we crystallographers do not despise anyone
: and we too use modelling techniques and tools as essential parts of
: our work. IMHO crystallographers are more closer to the truths
: of a biomolecule than anyone else. From my small experience I can
: say that a distance of 0.2 A between two non bonded atoms out of
: 13,000 atoms makes the difference between whether I have solved
: the structure or not. I doubt if modellers have scratched their
: head and spent two months about such a small measure that stands
: between truth and a model-protein which won't convert glucose into
: glycogen.
Please, don't get offended !, you are going over the same confusion again.
The level of certanty is a relative measure when it comes to protein
structures (or to anything in life, I suppose). A quantum chemist may
despise a 2 A resolution structure as useless, a cell biologist might just
have enough with a 10 A resolution receptor shape. Nobody is wrong or
right. What I object to is the usual assumption that modeling is a
useless exercise.
Cheers, Lluis
