From: Neil Errington <neil.errington@nottingham.ac.uk>
  To  : Gunther.Kern@astrazeneca.com, rasmb@alpha.bbri.org
  Date: Mon, 06 Aug 2001 09:46:59 +0100

Re: wiggles in dc/dt plots

Dear Gunther (et al)

I have recently observed the same phenomenon from with low concentration
experiments in the XL-I (never in the XL-A) and it made analysis somewhat
awkward but possible using DCDT. The "wiggles" in the residuals were
typically 1E-4 - 1E-6 fringes. I tried other methods with some success, such
as Peter Schuck's direct methods in Sedfit.

As to the problem itself, I'm not sure what causes it and maybe someone else
could provide an insight here..

Is it a window problem? It doesn't get  much better when running blank
corrections as compared to uncorrected data. But the blanks I have run have
all been "real time" blanks, in a different cell during the same run.

Any suggestions as to what this is and how we can get round it would be
greatly appreciated.

Neil

> <<10ug_ml BSA.bmp>>  <<Wiggles in AUZ VEL.bmp>> 
>
>I would be interested if anyone could comment on the wiggles I observe
>regularily when I use DCDT to analyze velocity runs. They are tolerable if
>the protein concentration is high enough as in the example I have included
>in my mail. However it totally renders the option to run at low protein
>concentrations since a wiggle can not be separated from what might be a
>signal. Quite a shame since DCDT to my experience on other machines allows
>analysis down to protein concentrations of 10 micrograms/mL. Here a few
>comments on how the experiment was run. It is a monomeric protein with a MW
>of 45 Kda dialysed extensively against 100mM Tris, 1 mM TCEP and 50 mM NaCl
>pH8. I also know from dynamic light scattering that the sample has a few
>percent of nonspecific aggregates which give the signal at around 6S. The
>speed was 50Krpm at 25 dgrees. Data collection was set up to collect
>interference data as fast as possible. The graph shows the overlay created
>by using 20 adjacent scans. Unfortunatelly I can not release the name of the
>protein but it is not a specific property of the protein. I do observe the
>same when using BSA (first bitmap).
>Any comment how to avaiod this problem would be very well appreciated. 
>
>Thank you very much,
>
>Gunther Kern
>AZ R&D Boston
>Infection
>+781- 839 4611
>gunther.kern@astrazeneca.com
>
>