Re: Time for CA-only models to be abolished ? (was: Re: [ccp4bb]: Howto deal with sidechainatoms with low electron density?)

[Edward Berry <eaberry@lbl.gov>]

***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***

Gerard DVD Kleywegt wrote:
> 
> Speaking of which - isn't it time for a grass-roots movement to root out that
> abomination that is CA-only models ?

No, I think there is a valid use for CA-only models - when the 
resolution is 4-5 A and the side chains don't show up but the 
transmembrane helices do, and the protein is of such importance 
that the biologists would prefer a TMH + Cofactors model
at 5 A to any number of mutant HEW lysozyme structrures at 1.2 A. 

2PPS (Photosystem I) was such a case and in my opinion one of the 
most exciting structures of the last decade. I believe some of the 
early cryo-EM structures of bacteriorhodopsin were C-a only, 
and were a huge improvement over the transmembrane-blob
structures we had been working with since 1977.  

It would have cut decades off the time it took for Ron Kaback
to solve the structure of lac permease (1M2U) by non-crystallographic,
non-NMR means, if he had had a good C-a trace available to
thread his structure onto.

And the ribosome- even more exciting- surely they don't have 
side-chains at that resolution. There is an immense amount of
useful information in a C-a trace, especially when it shows the
juxtaposition of cofactors or nucleic acids and the backbone.
For a lot of biologists, the first thing to do when one looks at 
a structure in RASMOL is to set the display to "backbone" or
"ribbons" to see how the thing folds. All those messy side 
chains just get in the way!

Ed