[RASMB] Weighted fitting residuals bigger than 1

[John Philo]

Hongjun,

When you do a weighted fit, I believe what is getting plotted on the
residuals plot is the ratio of the residual to the rms error for that data
point (as recorded in the absorbance scan file). Thus essentially the units
are the number of standard deviations, not OD. Somewhere in the output it
should tell you the rms error in OD.

John Philo
Alliance Protein Laboratories

-----Original Message-----
From: rasmb-admin at server1.bbri.org [mailto:rasmb-admin at server1.bbri.org] On
Behalf Of Jin, Hongjun
Sent: Friday, July 02, 2004 1:58 PM
To: rasmb at server1.bbri.org
Subject: [RASMB] Weighted fitting residuals bigger than 1


----------------------------------------------------------------------------
------
The older archived RASMB emails can be found at:
http://rasmb-email.bbri.org/rasmb_archives
and current archives at http://rasmb-email.bbri.org/pipermail/rasmb/
Search All the Archives at: http://rasmb-email.bbri.org/rasmb_search.html
----------------------------------------------------------------------------
------

Dear Collegues,

We are now using Bekeman Origin 3.78 to analysis our sedemetation 
equilibruem data, but we have a question regarding the high residuals of the

weighted fitting. We all know the OD absorbance usually is good around 0.1-
1.0, but why do we always get the range of the residuals over 1.0 or -1.0? 
The randomness distribution seem randomly well, but when I check the 
published papers, they normally have a residuals rangle around 0.02~ -0.02. 
WhenI use non-weighted fitting, the same data will give me a pretty low 
residuals range, but I do not have the 95% confidence intervals, which is 
not accurate from the previous discussions in RASMB.

Does anybody have similar experiences or know how to explain this, please 
let us know! Thanks  a lot!

Hongjun Jin

Department of biochemistry & Biophysics
Texas A&M University, College Station, TX 77840
Bio/bio Rm 237
lab phone: (979)-845-5617

email: hongjun-jin at tamu.edu





_______________________________________________
RASMB mailing list
RASMB at rasmb-email.bbri.org
http://rasmb-email.bbri.org/mailman/listinfo/rasmb