[RASMB] Strange AUC noise

Wed Oct 21 08:28:09 PDT 2015

David,

Regarding filling the cell all the way up to the top, it isn't just a number
of data points issue, it's primarily a separation issue.  The longer the
solution column, the more physical separation of different sized aggregates
you will achieve in your samples.  This is because separation is
proportional to time while diffusional spread is proportional to the square
root of time - longer solution columns facilitate increased total
separation.  Having that extra separation improves the ability to resolve
aggregates much more than the simple increase in number of data points.
This is particularly important if you are trying to separate an irreversible
dimer aggregate from the monomer.

Also, I really don't see any reason to worry about non-linearity in
pipetting for filling the cell.  We simply dial our p200 to about 230 uL and
add this volume twice, and this gets us all the way up to the top of the
solution column.  Who cares if 228 uL are delivered instead of 230 uL; we
only care about consistent loading and getting as much in the solution
column as possible.  Of course, accurate pipetting is paramount for
dilutions, but that's not what we are considering here... 

Also, if you run into leaking issues during loading - due to capillary
action drawing the liquid out of the cells - in my experience this arises
primarily from wetting the inside of the tiny filling holes from multiple
insertions of the pipette tip.  To avoid this, my standard loading procedure
is to insert a free gel loading tip into the filling hole and leave it there
- this keeps the filling hole dry.  I then put another gel loading tip onto
my p200 and add the solution by inserting the tip attached to the p200 into
the free tip I already placed into the filling hole.  Gentle pressure makes
a nice seal that allows you to pipette the contents of the top tip through
the bottom tip into the interior of the cell.  To avoid trying to pipette
against a vacuum, I lift the bottom tip up a bit with one hand, then pipette
with the other.  In this way you can load into the bottom tip as many times
as you want and not wet the inside of the filling hole.  The other important
issue is to load slowly to try to avoid the sides of the solution "crawling
up the windows" as the sample is loaded due to capillary action. This
procedure is especially useful for loading solutions that contain high
concentrations of detergents.

Regards,

Karl

  _____  

From: RASMB [mailto:rasmb-bounces at list.rasmb.org] On Behalf Of David Hayes
Sent: Tuesday, October 20, 2015 6:03 PM
To: RASMB
Subject: Re: [RASMB] Strange AUC noise

Hi Tom and all,

The practice in our lab is to use Spin50 cells always for absorbance
scanning since even at best it seems that Spin60 cells always have a higher
RMSD for absorbance scanning (maybe a bit less signal to noise due to being
narrower, or because of the timing issues, don't know, just observed).
Since we had been averaging about 600 samples a year, we have other
practices not found in the best academic labs.  I instruct my analysts to
load 400 microLiters on the sample side and 405 - 410 on the reference side.
We always use gel loading pipettes for loading to avoid scratching the walls
with needles.  We use Beckman supplied thin tubing or pipette tips to remove
the sample rather than blunt needles, also to avoid wear and tear on the
centerpieces.  Given that some samples are precious and we are always given
only enough for one filling, and some samples are irreplaceable, trying to
fill up closer to the top risks over filling spills both by getting the
bubble too close to the top and by dipping the pipette tip in three times
since most of our pipettes get a bit non-linear over 200 micro liters.   And
since the main data we want is percent aggregate, and we want to avoid as
much as possible any possible artifacts, we use absorbance data only and are
very conservative in choosing the area to fit fit.  Empirically I am of the
opinion that there are plenty of data points, and another mm of data is not
going to be very important to the SEDFIT final answer.   Jack has mentioned
to me before, and stated in this email chain, that his practice is to insist
on analysts optimizing the column height in the cells to near the max
possible, which makes sense theoretically, but practically I don't think the
extra data points change the final answers enough to justify the extra
stress and time taken by the analysts to try to achieve that optimization.
I get much more bang for the nagging by telling them to be more careful
washing their cells:  I have observed noticeable differences in final
answers between dirty and clean cells when doing replicates, and also when
the lamp is cleaned, I sometimes resolve peaks better in certain samples.

At first the Beckman repair person did not see anything wrong with the slit
assembly, but later after I wrote the email to RASMB the instrument failed a
radial calibration and he did replace the slit assembly.  Subsequently, the
instrument is now running OK again.

However, the last time we saw noise like this, replacing the slit assembly
helped for only a few weeks and then the noise came back, so I am not
completely convinced there is nothing else going on  (The last time we saw
this kind of noise, quite a few parts were changed for reasons that are a
bit complex and can't be put in the email).  The noise is never seen in the
scan of a blank hole without a cell, so maybe it is both the slit assembly
and something with the timing and cell light beam geometry.  We also had a
slit assembly become unglued in our oldest centrifuge and give the "jumping
pattern" noise John Philo described (as well as finding the bottom of the
cell to be outside the normal 7.15 to 7.2 cm range) but none of our older
machines have made this exact strange pattern.

Thanks for the discussion.  If there are any new developments, I will update
the conversation.  For now, changing the slit assembly has fixed the problem
and prevents further investigation:  an unexpected good news, since last
time changing out the assembly provided only limited help.

Kind Regards,

David

  _____  

From: "Laue, Thomas" <Tom.Laue at unh.edu>
To: "jphilo at mailway.com" <jphilo at mailway.com>; 'David Hayes'
<drdavidbhayes at yahoo.com>; 'RASMB' <rasmb at rasmb.org> 
Sent: Tuesday, October 20, 2015 8:54 AM
Subject: RE: [RASMB] Strange AUC noise

Hi-
Were these Spin60 or Spin50 cells? If they were Spin60, it might be that the
lamp timing on the 'noisy' machine is causing the light pulse to partially
hit the centerpiece wall. This is a machine-to-machine problem, consistent
with what you are seeing. However, the telltale for the timing problem is
increased noise in the scans, often at the top of the cell, not a wave.
Consequently, I vote for the slit assembly lifting (John's suggestion).
Tom

From: RASMB [mailto:rasmb-bounces at list.rasmb.org] On Behalf Of John Philo
Sent: Monday, October 19, 2015 5:00 PM
To: 'David Hayes'; 'RASMB'
Subject: Re: [RASMB] Strange AUC noise

David, I will speculate that you have a problem with the slit assembly. You
might see something like that if the foil on the bottom side has come
partially loose. Or perhaps there is some dirt or a burr causing the whole
assembly to lift up at certain radial positions.
I once had the foil come loose in a way that produced wild outlier scans,
but only every once in while (a few per run). Intermittently the foil would
actually catch on something (probably the ring on top of the PMT) and fold
completely back on itself.
And by the way, you should remind your people to fill the cells much more.
You want the meniscus to be up near 5.9 cm, not at ~6.25 cm.
John
From: RASMB [mailto:rasmb-bounces at list.rasmb.org] On Behalf Of David Hayes
Sent: Monday, October 19, 2015 11:14 AM
To: RASMB <rasmb at rasmb.org>
Subject: [RASMB] Strange AUC noise

Hi all,

We had some strange AUC noise in a recent experiment.  In the pictures below
from SEDFIT, the first two runs produced data as in the picture NNBekman1
(the IT guys name our instruments and they pruposely mispelled Beckman as
Bekman years ago). A strange noise pattern showed up in the scans that were
produced on day one.  Then the sample was loaded into a new cell and run
again on the same centrifuge producing the strange noise pattern in the
picture called NNBekman1.   the filled cell was taken and gently mixed and
then run on another centrifuge Bekman17 and made a perfectly fine patter
shown in picture Beckman17.

We have seen the strange noise before on our NNBekman1 centrifuge, but this
is the first time on NNBekman2.

Has anyone seen noise like this?  
Any theories on the cause of the noise?

It certainly looks like an optical artifact of some kind, but the time
dependence is hard to explain.

Kind Regards,

David Hayes

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.rasmb.org/pipermail/rasmb-rasmb.org/attachments/20151021/dc05cc73/attachment-0001.htm>