[RASMB] more on systematic differences

[Jack Kornblatt]

A follow up on "systematic differences"
There were two questions:
1.  If there are three cells in the AUC during a sedimentation velocity 
run, one cell yields an s value that I arbitrarily set at 100, cell 2 
yields a value of 99.4 and cell 3 a value of 99.7. What are possible 
causes of the differences?
2.  The meniscus of cell one if always found at shorter radius than the 
meniscus of cell 2.  Cell 3 is intermediate between the other two.  What 
are possible causes?

Thanks to John, Andrew and Tom for their most recent comments.
One of John's struck me as being especially insightful:

If you (me, in this case) have found that the three cells do not yield 
the same s values in a velocity run, and if you apply correction factors 
to two of them so as to get them to agree with the third, how do you 
know which is the correct one to use as a the reference value?  
The answer: You do not know which is the correct value.  If I understand 
correctly, to have knowledge of the s value to a precision of  0.1%, you 
would have to have experimentally determined diffusion coefficients (to 
better than 0.1% precision) as well as rho and vbar values to much 
better than 0.1%; mass spectroscopy now gives us molecular masses to 
high precision. Correction factors of less than 1% are useful in that 
they allow me to judge the sedimentation in one cell relative to the 
others.  Correction factors permit one to level the playing field but 
one doesn't know whether the game being played is football in San 
Francisco or cricket in Nairobi.
Like biologists all over, I had hoped for absolute knowledge when I 
started working with the AUC but ........

There seems to be general agreement that the differences between cells 
are probably the result of differences in the center pieces:  
I can easily check for leaks. 

NB. If I run one chamber empty, one full at 40 KRPM, am I likely to 
deform the center piece?

As for the other ideas as why the cells might differ as much as they do, 
they are all useful.  If the differences between cells were larger than 
it is, it would be worthwhile delving more deeply. The major unanswered 
question is why the s values differ.  Here you three all seem to agree 
that it might result from convection or from a leak.  I'll check the 
archives for how to detect convection problems and will get back to you 
about leaks when I have the data.  (Hopefully < 1 year)
thanks again
jack
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.rasmb.org/pipermail/rasmb-rasmb.org/attachments/20060906/6149a250/attachment.htm>