[RASMB] Data analysis with Sedfit: Smin setting affect the sedimentation profile and how we should deal with it?

[John Philo]

Yi-Ming,

Those results don't make much sense to me. Are you sure you are fitting the
same region of the data, and were these perhaps using fairly different f/f0
values? Notice that the first fit (starting from 0.2 S) has a significantly
lower rms residual. Taken at face value, that suggests that you really do
need to include some slowly-sedimenting material to explain your data (you
haven't shown us the region below 3 S so it's hard to judge what the fitter
wants to put there). But offhand I don't see why omitting the slowly
sedimenting fraction would lead to such big differences in the 10-20 S
range.

Certainly I find that many antibody samples do contain some fragments that
sediment at 1-2 S. Further, if your samples contain Tween-20 or Tween-80
then you will likely see sedimentation of micelles in that region too. So
you may well really need to allow for species below 2 S to explain your
data.

One key thing you have not told us is whether either of these distributions
actually produces a good fit of the raw data (low systematic residuals). If
neither is a good fit, neither will be reliable with regard to amounts of
minor components.

John Philo

-----Original Message-----
From: rasmb-admin at server1.bbri.org [mailto:rasmb-admin at server1.bbri.org] On
Behalf Of Yi-Ming_Li at hgsi.com
Sent: Tuesday, April 05, 2005 3:39 PM
To: rasmb at server1.bbri.org
Subject: [RASMB] Data analysis with Sedfit: Smin setting affect the
sedimentation profile and how we should deal with it?


Hi Everyone:

I am using analytical ultracentrifugation for quantitative detection of
aggregates in purified antibodies. Our experiments showed that the antibody
is not self-associated in the buffer system used for velocity sedimentation.
The molecular weight of the antibody is about 150 kd. The sedimentation
conditions are as follows:

Rotor speed: 40000 rpm
Temperature: 20C
Running time: 6 hours

When I used Sedfit to analyze the AUC data, I found that S(min) setting
affect the sedimentation profile quite significantly. I first set the
S(min) to 0.2 s and S(max) to 20 s. With this setting, I got a profile with
96.8% of monomer, 1.8% of dimer and 1.4% of species larger than dimer as
showed in panel A in the figure. Since no peak was detected between 0.2-2 s
(although a spike of the base line at the Smin end was found), I reloaded
the data, and set the sedimentation coefficient range to 2-20 s for data
process. This time I got a profile with 98.9% monomer, 1.1% dimer, and no
species larger than dimer as showed in panel B in the figure. Moreover, the
sedimentation coefficient of the dimer is different in the two profile (8.9
s and 9.7 s).

It is important for us to know: 1) Is the way I set the S range right? 2)
Are the species larger than dimer found in panel A true species or just
artificial? 3) why the Smin setting affect the profile so much?

I would appreciate very much for any commend or discussion.

Thanks!

Yi-Ming

(Embedded image moved to file: pic05196.pcx)