From: John Correia <jcorreia@biochem.umsmed.edu>
  To  : rasmb@alpha.bbri.org
  Date: Fri, 30 Apr 1999 11:59:10 -0600

my 3 cents

What is not being conveyed in these responses to the question about the
reliability of D is the actual precision that one might expect from each
technique.  5%? 10%?? 25%???  

It is my guess that each of these approaches (Svedberg, s/D from DCDT and
finite difference and maybe vanHolde Weischet) will give nearly the same
answer +/- if the sample actually is a single, ideal noninteracting
species and the method is used properly.  So the fact that everybody is
promoting their favorite method, the one they wrote or developed, the one
they primarily use, doesn't convey the fact that each of them works for
good samples.  

That is not the real problem.  The issue is what if the MW is incorrect -
how do you sort out the cause and worse than that how do you fix it, make
it go away?  This proves to be in my opinion a much harder and more
interesting question since it really depends upon the extent of
heterogeneity or aggregation or association or all three in the sample.  It might
even be instrumental or the way you collect data, the speed & the number
of scans used, the concentration, the presence of nonideality.  Is it XLA
or XLI data?  Once again I suspect each method, with caveats, could be
used to sort out the problem.  

We have observed cases where MW by sed equil is on the money but s/D is
low, sometimes up to 50% low.  Sorting it out is proving to be educational
& challenging.  Rather than tell you what we are doing, what my bias is, I
recommend comparing methods for your particular case or sample, think
about each method and try to understand what each is telling you.

-------------------------------------------------------------------
 Dr. John J. "Jack" Correia
 Department of Biochemistry
 University of Mississippi Medical Center
 2500 North State Street
 Jackson, MS  39216
 (601) 984-1522                                 
 fax (601) 984-1501                             
 new email address: jcorreia@biochem.umsmed.edu     
 homepage: http://fiona.umsmed.edu/~biochem/correia.html  
-------------------------------------------------------------------