Documentation

List of publications related to the development and applications of HDBStat!.

• Prinal Trivedi, Jode W Edwards, Jelai Wang, Gary L Gadbury, Vinodh Srinivasasainagendra, Stanislav O Zakharkin, Kyoungmi Kim, Tapan Mehta, Jacob PL Brand, Amit Patki, Grier P Page and David B Allison. HDBStat!: A platform-independent software suite for statistical analysis of high dimensional biology data,BMC Bioinformatics 2005, 6:86
  HTML, PDF

Two different input files, gene expression data and chip level information are required to be able to use this software.

• Gene expression data file: First column MUST be probe names and first row MUST contain chip names starting from second column. All other cells MUST contain numeric data specifying gene expression levels.
• Chip level info file: First column must be chip names and first row MUST contain class variable names, which include treatment name(s) and applicable pairing name(s) starting from second column.
• If you have two-color data, you can use our two-color data management tool to produce HDBStat!-ready input files.

• Dataset 1
  • Gene Expression Data File
  • Chip Level Info File
    
    
• Dataset 2 (paired)
  • Gene Expression Data File
  • Chip Level Info File
    

1. Set Up I/O
   • User Input: Select two input files - gene expression levels and chip level info file. Select a directory where you want HDBStat! to output results
     
     
2. Preview Data
   • User Input:
     In the "Chip level info" tab, select a set of class variable(s) in "Select Class Variable(s)" window, and pairing variable in "Select Pairing Variable" window (required only if analyzing paired data).
     
     
   • Results Displayed:
     After gene expression data and chip level data are imported, they are displayed in "Gene expression data" and "Chip level info" tabs respectively. Chip level statistics such as Pearson? correlations between chips, mean, standard deviation, minimum and maximum values of gene expression levels are calculated on raw data and displayed in graphical and tabular format in different tabs.
     
     
3. Preprocess Data:
   • User Input:
     This step is optional. If you wish to apply preprocessing, select a method. If you select a method that also involves transformation, select a log base under "Options" section. Click on "Apply" button.
     
     
   • Results Displayed:
     If a preprocessing method has been selected, it is applied to the raw data and again, chip level statistics for preprocessed data is displayed in tabular and graphical format in different tabs.
     
     
4. Analysis:
   • User Input:
     In this step, you will set up hypotheses, select a method of quality control and select testing methods. Based on the variables selected in step 2 and information provided in chip level information file, a combination of all possible hypotheses is displayed in "Set up hypotheses" step. Select at least one hypothesis in order to perform a two group comparison. As a quality control measure, you may choose to perform Deleted Residuals in "Select quality control" tab. In "Select tests and multiplicity control" tab, you can select multiple methods of hypotheses testing and multiple comparison adjustments
     
     
   • Results Output:
     Results in form of CSV, MS Excel and image files from quality control and hypothesis testing methods are output to the user specified directory. HDBStat also generates a HTML file that provides a summary of the analysis. This mechanism of outputting results provides the user an opportunity to view quality control results and modify hypotheses, preprocessing method, and/or chip selection before proceeding to the next step.
     
     

• Dataset 1 (11MB .AVI file - best played in Windows Media Player)
• Dataset 2 (11MB .AVI file - best played in Windows Media Player)
• Dataset 3 (13MB .AVI file - best played in Windows Media Player)

• Click here

• PDF