From: Matthew Parker <parker@topaz.microbio.uab.edu>
  To  : rasmb@bbri.harvard.edu
  Date: Fri, 30 May 1997 12:06:47 -0500

Confidence limits for nonweighted fits using Origin

Hi! We have been using the Origin package which comes with the Beckman
software to calculate Ka's from equilibrium data. I am happy with it,
except for one thing. When you do a nonweighted fit using multiple
channels, the confidence limits for the fitted values are not reported.
If you do a weighted fit, it does give confidence limits, but the
problem is that the plot of residuals for weighted fits does not use
absorbance units in the y-axis (and so it's not really a
publication-quality plot).

	Does anyone know if Beckman or MicroCal have come up with a patch or an
upgrade which solves either of these problems? And, while I'm on the
subject, which do you recommend: weighted or nonweighted fits?

	Thanks!

	Matthew Parker