Custom Query (361 matches)


Show under each result:

Results (22 - 24 of 361)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Ticket Resolution Summary Owner Reporter
#348 fixed ga_init statistics errors gegorbet demeler

There are problems with the statistics calculation in ga_init that need to be checked into it. Example:

Run ID: 105v1-2 Cell 2, Channel B, Wavelength 280 Edited Dataset: 1205151550

Solute 2: Molecular weight: 4.9746e+04 (5.0392e+04, 5.8040e+04) Sedimentation coefficient: 2.1050e-13 (2.0976e-13, 2.1224e-13) Diffusion coefficient: 3.7237e-07 (3.1523e-07, 3.6712e-07) Frictional ratio: 2.3745e+00 (2.3958e+00, 2.6522e+00) Partial concentration: 4.5759e-02

The values in bold do not make sense, the mean is outside of the confidence interval. Something is calculated incorrectly here.

Another problem with the statistics page is that the percentages are not calculated correctly. What should be done is the following:

The total concentration of the experiment (all solute concentrations summed up will equal the total concentration of the experiment) corresponds to 100%. Any bucket selected will be a fraction of this total. So if you have say 6 groups, one of which is bracketed by the user, than the percentage of this bracketed solute can never be 100%, yet it is listed as such if only 1 bucket is selected. The only time this can be 100% is if only one group is there and all of the solutes are bracketed in a single bucket. So, instead of calculating the percentage of just the bracketed solutes, this calculation needs to be done in reference to the entire solute concentration.

#347 fixed dcdt reports gegorbet demeler

A small change: The labeling of the distributions derived from the dcdt analysis should be called g*(s), not G*(s). A capital letter implies an integral distribution (i.e., van Holde - Weischet), while a lower case implies a differential distribution (ie., dcdt).

#346 fixed 2 additional pseudo-3d pre-filter combine issues gegorbet demeler
  1. Clicking on "Reset" does not reset the runID prefilter
  1. For admins, it is not possible to choose the data from a different user in the pre-filter selection window. Right now, one would have to choose the user through "Load Distribution" first, which defeats the purpose of the pre-filter. I think adding a user-selection widget to the prefilter dialog will take care of this.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Note: See TracQuery for help on using queries.