[RASMB] Interference Optics - HELP!

[Jeff Lary]

Lisa,
Your questions actually deal with a number of different issues that are
involved in setting up and using the interference optical system and there
is no short answer but I'll try to be as brief as possible.

In my opinion, the tilt of the fringes as seen in an image of either an
empty rotor hole or the rotor scallop should be removed by rotating the
camera to give you a level interference pattern.  The major problem with not
doing this is the possibility that when transforming the fringe data in a
region of a steep gradient you would be using information from different
radial positions to calculate the phase of the transform.  In effect you
would be either decreasing or increasing the apparent gradient depending
upon whether the baseline tilt was up or down (I think that's the way it
would go).
Performing a blank subtraction - and I mean a blank that is taken through
either an empty rotor hole or the scallop, not a saved blank file - will
only remove the tilt from your data and not correct for any error introduced
as described above.
I don't believe that there is any post-processing that can be done that will
correct for the error that may be introduced by transforming data that has a
significant "baseline" tilt.  On the other hand, I also can't tell you what
the magnitude of the errors would be since I've never taken an interference
pattern of an experiment where there is a high gradient, transformed the
pattern through rotation and then done the Fourier transform on the rotated
data.
Now onto the next problem - actual blank subtraction.  We have some cells
(I'm talking here about 6-channel cells for sedimentation equilibrium
experiments) where the water/water "blank" data is essentially flat and
level for each of the three pairs of channels and so doing a blank
subtraction wouldn't have a significant effect on the quality of the data.
On the other hand we have some cells where, for the water/water blank, the
top channel may be flat, the middle channel may slope upward (or be curved)
and the bottom channel may slope downward.  In this case the subtraction of
a water/water blank will have a significant effect on the quality of the
data.  If you are using 6-channel cells then they should be of the
external-loading type so that you can obtain the water/water blank without
disassembly.  If you are using the standard double-sector cells then you can
assemble the cells, fill them to the appropriate height with water in each
sector, take them up to the speed (or do this at each speed for multiple
speeds) and take a set of data and use this to blank-subtract the data from
the actual run.  Then empty the cells and vacuum dry them, fill them with
your sample & buffer and do your run.   Though I've never used the choice in
the XL-I program to use a blank file, it looks like it should work if you
rename the water/water blanks as "blank.ip1", "blank.ip2", etc.  Other than
that you would have to do the subtraction outside of the XL-I program.
As Kristian noted, I would rather do this type of blank-subtraction after
the experiment because, in effect, you are altering the "pseudo-primary"
data by doing this subtraction during the data acquisition.  I say
"pseudo-primary" (and please excuse me if I go off on a tangent) because the
primary data, the actual image intensity data, is always discarded by the
Beckman software.

I hope that some part of this may help,

Jeffrey Lary
National Analytical Ultracentrifugation Facility
Biotechnology Center Unit-3149
University of Connecticut
ph: 860-486-5036
fax: 860-486-5005
email: jeffrey.lary at uconn.edu
===========================

----- Original Message -----
From: Lisa Joss
To: rasmb at rasmb-email.bbri.org
Sent: Sunday, March 16, 2003 10:21 PM
Subject: [RASMB] Interference Optics - HELP!


Hello All,
We have recently purchased an XLI and I have some questions regarding the
interference optics that I am hoping some of you can help me with. My
problem is basically non-linearity of the scans, and correction of that
non-linearity using a blank scan correction.

Firstly, when scanning an empty rotor hole the slope of the best fit line to
the resulting data is about 0.2 finges/cm (well outside Beckman's specs of
0.04 fringes/cm). Scan's of an empty cell are comparable.  Beckman is not
terribly concerned about this and will have a service rep align the camera
to try and improve, they suggest that if I just do the appropriate blank
scan subtraction everything will be OK (confidence that I do not share!).

Which brings me to my next, more troubling, concern - blank scan
subtractions. Subtraction of a blank scan taken of the same empty rotor
hole, empty cell or water cell does not always give me a flat baseline,
particularly when I use the option of subtracting a previously saved scan.
When dealing with only one cell the result is fine, however when I do
several cells in one scan only the one whose baseline was scanned last is
flat, all the rest have a significant deviation from zero. I think what's
happening is that even though all cells have a blank file (blank.ipx) all
the subtractions are using the last file scanned. We have version 4.5 of the
data acquisition software and it doesn't have an option to select an actual
scan so I'm assuming that it should just use whatever the last blank.ipx is
for the relevant cell. I have tried deleting some of these blank files and
the software is looking for them - apparently just not using them. Can
anyone shed any light on this?

While I agree that I need not worry about this non-linearity for our g*(s)
analysis, I am trying to get good equilibrium data for a system that we have
to use interference for (buffer contains nucleotide) and I am sure this is
causing a problem. Any suggestions or information would be appreciated.

Thanks

Lisa




Lisa Joss                                                            (801)
585-3919
Postdoctoral Fellow                                           email:
ljoss at biochem.utah.edu

UNIVERSITY OF UTAHBIOCHEMISTRY20 N 1900 E RM 211SALT LAKE CITY UT 84132-3201