Functionality Requests

  • DONE Change T-test p-value calculation from equal variance to unequal variance.
  • Eisen-like data overview with sortable columns and groupings.
    • Ability to "crop in" to Eisen view.
  • Automatically display first few selected patients in missed column view
  • Include a "images" folder, save all images to this folder.
    • Would it make more sense to put the images in the folder with the distance output? Only if we want to output more than just the summary chart (like clusters). The point of having the images folder is to keep them all centralized.
  • DONE Include a "cluster" folder to store cluster output (not images).
  • Automatically save distance result graph and resave when points are selected.
  • Automatically name graphs "graph_" + "critmanager_traingp_testgp" + ".png", or something better.
  • Automatically save cluster graph with defaults in the images with a informative name.
  • DONE When saving a cluster, suggest or default a minimum size that will allow the cluster to be saved without any overlap on the labels.
    • If size is too small, labels are automatically disabled.
  • DONE Default cluster to: Include Spectra, with labels and with group colaration.
  • DONE When drawing the cluster in the interface, show the group coloration.
  • DONE When drawing/saving a topN graph, check the length of the names in the legend. Adjust for the lenght of the text, so that text does not overlap.
  • When drawing/saving a topN graph, check the length of the percent values on the left hand side, so that the values are all visable.
  • DONE When generating clusters, automatically select the points that have been selected on the main graph.
  • DONE Ability to remove a Distance Run from the tree.
  • DONE Ability to remove a cluster graphic from the tree.
  • Mass Save of images via a single command.
  • DONE Set cluster dimensions based on the square pixel size.
  • DONE Set color classification bar wide enough to fit all labels.
  • Prompt the user for what to do when averaging data: ignore 0 (MassSpec) or not (MicroArray).
    • Maybe incorporate this into program start/dataset load. Ask the user what the data type is and store in project data or dataset info. We could make the 2 log buttons into a single one and decide the base (2/10) by the data type. That would also determine how to handle 0 for average.
    • This falls under the data properties enhancements below.
  • We could also determine the data type by the mass/name column for the most part.
  • Provide filtering in the DataSet and Score views. See:
  • Generate pre-post datasets. (similar to average) See Dr. Shyr for actual methods.
  • DONE Re-generate score summary after adding external score file?
  • DONE Generate ROC for testing data and include in project data
  • Fix ROC assumption that cases are group 1 and controls are group 2
    • Temporary hack is at Roc.cs: line 225.
  • VizSpec options:
    • Select curve color, order, offset
    • Edit legend


  • Add paired test functionality for T-test and SAM for pre & post datasets.

Data Properties Enhancements

  • ProjectData or DataSetInfo
    • designate what type of data (micro array, mass spec) and/or the log base (2, 10) - or this could be in DataSet itself
  • Score
    • DONE designate which dataset should be used (raw, log, either).

Low Prorioty Enhancements

  • DONE Ability to save and load CriteriaManager (at least CutoffManager and see below) from file (for non-release version only?). -- This will help on large projects when the same CutoffManager criteria is used for different DataSets.

Topic revision: r23 - 03 Mar 2006, WillGray

This site is powered by FoswikiCopyright © 2013-2022 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Vanderbilt Biostatistics Wiki? Send feedback