Yes the BV method relies on summing up the 'bond valence' contributions for each ligand co-ordinated to the metal, and comparing the result with the formal charge on the metal ion (e.g. +2 for Mg). This means that even if say one water that is present in the crystal is omitted from the calculation for whatever reason (say you summed the contributions from only 5 instead of 6 waters because you didn't include the symmetry molecule as in the original version of WASP), then you are likely to conclude from the result that it's not Mg even if the distances and the octahedral co-ordination of the 5 waters that you did include are perfect for Mg.
Obviously as George says, if the distances are inaccurate then the computed BV contributions to the sum will be wrong and that's not good either! So I completely agree that it's probably wise to take the results of applying this method with a large pinch of salt! Cheers -- Ian On Tue, Dec 21, 2010 at 12:25 PM, George M. Sheldrick <gshe...@shelx.uni-ac.gwdg.de> wrote: > > You might also like to look at the paper "Is the bond-valence method able > to identify metal ions in protein structures", Acta Cryst. D59 (2003) > 32-37. In retrospect, the BV method is very good for identifying Mg2+ > because it is almost always tightly octahedrally coordinated by oxygen > atoms. For other metals and low resolution the calculation can easily be > upset by missing water ligands or inaccurate interatomic distances. > If anyone is interested I can supply a powerpoint of the lecture on the > BV method from my undergraduate solid state chemistry course, but I > should warn you that it is in German (like most of our teaching). > > George > > Prof. George M. Sheldrick FRS > Dept. Structural Chemistry, > University of Goettingen, > Tammannstr. 4, > D37077 Goettingen, Germany > Tel. +49-551-39-3021 or -3068 > Fax. +49-551-39-22582 > > > On Tue, 21 Dec 2010, Ian Tickle wrote: > >> Note that the original Nayal & Di Cera algorithm uses an older and >> less preferred version of the bond valence model and its associated >> parameters than current implementations of BV. >> >> The original WASP uses this formula: >> >> bond valence = (Rij/R0)^(-N) >> >> The most up-to-date BV parameter set >> (http://www.ccp14.ac.uk/ccp/web-mirrors/i_d_brown/bond_valence_param/bvparm2006.cif >> ) uses this: >> >> bond valence = exp((Ro-Rij)/B) >> >> where Rij is the metal-ligand distance, and R0, N and B are parameters >> specific to the metal-ligand pair. Obviously these parameter sets may >> give different results depending on which formula you are using. >> >> See the IUCr monograph by I.D.Brown: "The Chemical Bond in Inorganic >> Chemistry: The Bond Valence Model", chapter 3 for more info. >> >> Cheers >> >> -- Ian >> >> On Tue, Dec 21, 2010 at 8:55 AM, Robbie Joosten >> <robbie_joos...@hotmail.com> wrote: >> > You could also try the original WASP here (also for coloured Indo-dutch >> > catholics): http://xray.bmc.uu.se/cgi-bin/gerard/rama_server.pl or you can >> > use the latest WHAT_CHECK that also has an implementation of Nayal and Di >> > Cera's algorithm. >> > >> > Cheers, >> > Robbie >> > >> >> Date: Mon, 20 Dec 2010 22:59:53 +0000 >> >> From: paul.ems...@bioch.ox.ac.uk >> >> Subject: Re: [ccp4bb] Mg2+ or water >> >> To: CCP4BB@JISCMAIL.AC.UK >> >> >> >> On 20/12/10 21:48, Robbie Joosten wrote: >> >> > >> >> > >> >> > Also note that Mg2+ is significantly smaller than water. It fits in >> >> > places where water cannot go. This doesn't look like a magnesium site >> >> > on first glance. >> >> >> >> I tend to agree with Robbie. I wonder what WASP would say... (if you >> >> use Coot, you can try the "Highly Coordinated Waters" validation test - >> >> a symmetry-enhanced implementation of the Nayal & Di Cera (1996) >> >> algorithm). >> >> >> >> Paul. >> > >> >>
