Up ] Load New Files ] Set Initial Data ] [ Data Housekeeping ] realtime ] Export to SEDPHAT ] Exit ]


Edit Data Files

Data | Edit Data Files

This function will start the notepad and allows loading and inspecting a data file in text mode. This allows changes, such as the manual elimination of spikes, reading the comment line, run parameters, etc.


Save Fits

Data | Save Fit Data

Fitted distributions, residuals and fitting parameters can be stored on the hard disk.

For some models (c(s), c(s,f), and c(s) or c(s,f) with discrete component) this automatically writes the ~tmppars.* file to the directory where the data are located, such that the current fit can be restored when loading the data.

First, the calculated best-fit radial distributions can be saved in files with filenames that can be specified in the ‘save fitted data as’ dialog box. The filenames of the fitted curves proposed depend on the filename convention that was specified (newer Windows Beckman software assumed as default). For the newer convention (Windows XLA software), the first character of the original data filename will be replaced with a ‘f’ - for ‘fitted’ and the rest of the name will be kept the same. This way, for each data file (e.g., 00006.ra2) a new file (f0006.ra2) is created.  For the older convention (old DOS software for 4-hole rotors), where the file extension identifies the scan sequence, the extensions are kept for the saved distributions, but filename can be changed (e.g., ‘fitcell1.r01’). Please note that it is currently not possible to deviate from this rule.  Therefore, a new directory should be chosen to save multiple alternative fits!

The saved data will contain the calculated best-fit distributions as shown between the green lines in the upper plot in Sedfit.  The files will be in XLA format and compatible to re-analysis with Sedfit or other software. Analogously, the residuals will be saved with a 'r' as first character in the filenames, and be saved in XLA format.  For the parameters, a file like ‘fitparams.ra2’ will be generated. Information saved in this file will include, for example,

component 1 included:
M = 65000.000000 g/Mol (floated)
s = 4.514 S (floated)

component 2 included:
M = 130000.000000 g/Mol (floated)
s = 6.510 S (floated)

component 3 included:
M = 200000.000000 g/Mol (floated)
s = 8.468 S (floated)

component 4 included:
M = 10000.000000 g/Mol (floated)
s = 2.005 S (floated)

concentrations = 2.194 | 0.108 | 0.041 | 0.025 | (first is total, next are fractions)

meniscus at 6.2299 cm (fixed)
bottom at 7.2206 cm (fixed)

baseline at 0.00000 OD (floated)
(time independently fitted)
fitted from 6.2532 to 7.1550 cm
with constant initial distribution
simulation start time 0.000000 sec (fixed)

rmsd=0.006015 (n=54309,SSR=1.965085)

If ‘cancel’ is pressed in one of the dialog boxes, the respective operation is skipped.

!    There is an alternative procedure for saving data, which involves copying the data and fits to the clipboard and pasting them into a spreadsheet program.  This has the advantage that they can also be plotted and saved in a single file.  Since the fit parameters are not complete, separate notes should be taken of the main results.   


Save Systematic Noise

Data | Save Systematic Noise | Save Noise in File

If the data analysis has been performed with the options for time-invariant (TI) noise (i.e. the systematic offset in radial direction superimposed to all scans) and/or the radial-invariant (RI) noise (i.e. a systematic noise component that causes an offset to each scan which is constant in radial direction, but different for each scan) , these files can be stored to the hard disk using this menu item. Filenames for the time-independent noise are "*.TIN", filenames for the radial-independent noise are "*.RIN". Time-invariant noise is saved for the data analysis range (between the green lines) only, and in XLA file format.  Radial-invariant noise is saved as simple two-column ASCII file, with elapsed seconds in first column RI offset in the second column.

Note that once the systematic noise has been subtracted from the data displayed in the sedfit window, any information about the original TI and RI noise is lost!  Therefore, this function should be used only when the true raw data are visible.  


Correct Equilibrium Scan for TI Noise

Data | Save Systematic Noise | Subtract Noise from Eq. Data

If the data analysis has been performed with the time-independent noise option, the calculated systematic TI noise can be subtracted from an equilibrium scan.

The strategy would be the following: First, load as set of scans from the approach to equilibrium (should cover a substantial part of the visible sedimentation process), and analyze them allowing for TI noise.  The model used is secondary, but it is of highest importance that an excellent fit is achieved without any systematic residuals.  Second, use this menu to load a scan from sedimentation equilibrium and subtract the previously calculated TI noise. Save the new file with a new name (suggested extension: is "*.RTI"). This new file contains equilibrium data corrected for the systematic noise contribution, as identified by the sequence of approach to equilibrium scans. This should improve the quality of the equilibrium profile and of their analysis. In particular, this can eliminate or reduce the errors in the equilibrium scans resulting from imperfections in the window. The radial grid points of the equilibrium data are not changed for the subtraction. The method seems to work best for low loading concentrations.  It has bee described in detail in 

P. Schuck (1999) Sedimentation equilibrium analysis of interference optical data by systematic noise decomposition. Analytical Biochemistry 272:199-208.  

The experimental procedure for eliminating the time-invariant noise contribution from equilibrium scans is the use of external loading cells and the acquisition of before and after-blanks. This is described in detail in 

Ansevin et al., Anal Biochem 1970, 237-261.

!    Great care should be taken that the TI noise component does not contain systematic patterns, resulting from a the use of an insufficient model in the approach to equilibrium analysis.  The quality of fit of the approach to equilibrium data should be optimal, and the use of a very large model is recommended, which not necessarily has to reflect properties of the true solutes.  After all, the modeling of the approach to equilibrium could be considered here an informed smoothing operation employing all possible solutions of the Lamm equation, without any further interpretation beyond the extraction of TI noise. For example, with absorbance data, fitting with a floating starting time was occasionally found to be very useful.  If the TI noise contains any systematic drift, this can introduce systematic bias into the new equilibrium data.  If the TI noise does not contain a systematic pattern or drift, but looks randomly distributed, it should provide a good correction.   

Note that once the systematic noise has been subtracted from the data displayed in the sedfit window, any information about the original TI and RI noise is lost!  Therefore, this function should be used only when the true raw data are visible.  



Save Calculated Distribution

Data | Save Continuous Distribution

This feature allows to save the calculated distribution (e.g. c(s), ls-g*(s), c(M), or van Holde-Weischet) to a file, as a two-column tab-delimited ASCII text.



Data | Print

Data | Page Setup

Only limited options for printing are currently available, which allows only to get a quick screenshot for documentation of the data, fit, and results.  For publication quality graphics, it is recommended to use the copy commands to place the data in the clipboard, and to paste the data into a spreadsheet of a commercial graphics program. With Microcal Origin, for example, one can setup worksheet and figure templates that allow quickly generating publication quality figures.

Using the print command, Sedfit can only print the contents of the current Sedfit Window (similar to a print screen operation). The position and size of the graphics on the paper will somewhat dependent on the position and size of the Sedfit Window in the desktop.  

For better results, the residuals or the data plot, respectively may be toggled off (see display commands). The printout can be made in color or black and white, and with or without the fit information section from the upper left corner of the Sedfit window.