[RASMB] XL-I poor data troubleshooting

[John Philo]

Mark,

You didn't explicitly say whether you are talking about absorbance or
interference data. One key diagnostic would be whether both optical systems
give similar problems. If both are messed up, I think it is likely you have
a severe temperature control problem (the instrument may think the
temperature is constant, but it isn't really). So if both types of data are
bad, I would first check the radiometer for oil on the plastic cover or
(even worse) covering the little black disk that absorbs the IR radiation.

However since you mentioned the radial servo it sounds like it is only the
absorbance data that is bad. If that is true I think most likely this new
radial servo is giving bad radius values. I am aware of a couple of
instruments where the radius readout was or became strongly non-linear (the
values were correct at the calibration points, but nowhere in between). One
of those was a brand new instrument so the readout potentiometers can be bad
when new. 

Unfortunately none of the standard Beckman field testing or QC tests will
detect such a problem. To test for it, try the following: Load an empty
6-channel centerpiece with no windows in the window holders. Take that up to
3000 rpm and record an equilibrium-mode INTENSITY scan. In that scan both
the sample and reference channels should show a square-wave type pattern
with zero intensity whenever the light is hitting the black epon. Thus the
edges of the square waves define where the instrument thinks channels A, B,
and C are located. The key is that the channels are actually all the same
length, and the ribs between the channels are the same thickness too. If the
readout is non-linear that will not be the case, and the center of channel B
will not be located at 6.500 cm as it should be.

If your instrument is truly imaging parts of the walls of the centerpiece
then you would get wrong OD readings for the standard OD test solutions that
come with the instrument, so you could look at that too, but I doubt this is
the problem.

With respect to the repetitive decimals on the sedimentation coefficients,
this is caused by the spacing of the data points as defined by the SEDFIT
parameters page. If your main peak is only one data point wide (the peak is
triangular), the weight-average sedimentation coefficient for that peak has
to be exactly one of the x-axis grid values. This is a common error---if you
are going to report a sedimentation coefficient from c(s) or ls-g(s) be
careful that the resolution is high enough and that the strength of the
normalization is high enough so that the peak is spread over several data
points.

John

-----Original Message-----
From: rasmb-bounces at rasmb.bbri.org [mailto:rasmb-bounces at rasmb.bbri.org] On
Behalf Of Mark Agacan
Sent: Monday, September 08, 2008 3:46 AM
To: RASMB at rasmb.bbri.org
Subject: [RASMB] XL-I poor data troubleshooting

Hello,

I am having problems with the data being collected on my XL-I.  For all the
samples I have studied lately, the rmsd has been 0.2 - 0.8, the residual bit
maps show large contrast from black to white running in a diagonal across
the whole data range, and the MW values output from SedFit bear no
resemblance to what I have been  expecting.    

I have also noticed that the S-values just don't seem right:  e.g. cell 2 =
6.545455; cell 4 = 7.272728; cell 6 = 8.383839; The repetition in the
numbers after the decimal point are surely artefactual?  I've never seen a
dataset consistently produce numbers like this.

I have used different windows and centerpieces in an attempt to get better
data but to no avail.  My sample preparation is standard and the samples and
buffers are standard, e.g.  50 mM Tris 100 mM NaCl.  All samples show up as
distinct bands on gels with MW values as expected.

I recently had the radial servo replaced on my XL-I - could this or
malfunctioning / poorly calibrated optics be to blame?  I will run a
lysozyme sample (MW = 14.4 kDa) in the next few days in place of a suitable
standard.  The residual bitmap and high rmsd makes me think that the optics
may not be properly aligned and that they are imaging part of the walls of
the centerpiece, instead of the sample.

It's not impossible that the samples actually really do have the mass and
S-values as output, but I am highly suspicious, and the rmsd values are
terrible.

Has anyone had issues such as this, and if so have you any recommendations
or ideas that I can try in order to get to the bottom of this problem?

Best Regards,

Mark


_____________________________________
Dr Mark Agacan
Scientific Officer,
Division of Biological Chemistry
and Drug Discovery,
Wellcome Trust Biocentre,
College of Life Sciences,
Dow St.,
University of Dundee,
Dundee, DD1 5EH
Tel: +44 1382 388751
Fax: +44 1382 345764
_____________________________________
The University of Dundee is a registered Scottish charity, No: SC015096
_______________________________________________
RASMB mailing list
RASMB at rasmb.bbri.org
http://rasmb.bbri.org/mailman/listinfo/rasmb