[RASMB] baseline, radius values, & variance

[McCornack, Melissa A]

Dear RASMBers,

I recently began performing sedimentation equilibrium experiments and have 
been trying to use mainly Winnonlin to analyze my data (have tried Origin to 
a lesser extent). In the process I have encountered some difficulties, some 
of which I found answers for in the archive and some not.  I am hoping that 
by writing to all of you maybe someone can give me some pointers.

1)My first question(s) have to do with the baseline offset.  I have read in 
many papers that experiments were conducted at angular speeds of greater 
than 40k rpm or that overspeeding was used to deplete the meniscus.  I 
understand that the value obtained would give a baseline correction term.  

What I am unclear on is where that number (~0.01 absorbance units in my     
experiments) is used.  

In Winnonlin, do I put it in for all the DeltaYs and then constrain them to 
that value (even though for experiments at lower rpm the meniscus is not 
depleted)? Is this even reasonable as the experiments at lower speeds are 
not depleted? Is it better to have the DeltaYs float but then why do so many 
papers say they found the value for the depleted meniscus?  Is there another 
place to input this number in Winnonlin so that it is subtracted from all 
experiments as a baseline correction term?  If yes, do we then allow the 
DeltaYs to float in the fitting?  Furthermore, I have found that when I just 
let the DeltaYs float that my experiments at 40k rpm fit well to the 
depleted value (0.01) but that those at speeds less than 40k the DeltaYs are 
on the order of -0.015.  Is a negative number anything to be alarmed at?

2)Along with question 1, how and when do we use the Special-Change Variable 
Offset function in Winnonlin?

3)I have been looking at the effect of editing the starting radius (r) value 
and the ending r value in my data sets.  In a set of experiments all run 
with the same concentration (i.e. same sample) but varying speeds, I edit in 
Winnonlin the starting and ending r values for all experiments so they are 
identical and find a Ka.  Then, I alter the starting and ending r values in 
all the experiments and find a new Ka.  Sometimes the Ka’s are comparable, 
sometimes they are not. 

I am wondering if the placement/editting of the starting and ending r values 
used in analysis are crucial. 

(Note: I am staying between 0.05 and 1.2 absorbance units as I read on the 
web somewhere the Beckman XLA in absorbance mode was linear to around 1.2).  
I have noticed that sometimes above 1.0 I get residuals that are very 
skewed.  Is it reasonable to edit that data out or not?

4) Finally, I was wondering if someone who has worked in this field awhile 
might be able to give me an idea of numbers to expect for “variance of 
fit”, “sum of residuals squared”, and “square root variance” for a good fit 
of good XLA data in Winnonlin. 

As I said I only recently started doing AUC experiment (3 weeks to be 
exact), I am grateful to all those who have contributed to the RASMB archive 
as it has already been very useful. Finally, let me say in advance thank you 
to all those who not only read this rather long series of questions but 
especially to those who take the time to response-thank you for you help.

Sincerely,
Melissa McCornack Ph.D. 
Texas A&M University
Department of Biochemistry and Biophysics
MS 2128
College Station, TX 77843-2128