User: Anonymous User ( Login | Register )

False Discovery Rate controlling procedures for multiple statistical test P values


Introduction
This is a HIGHLY experimental resource which implements a variety of procedures for controlling the Type 1 error False Discovery Rate when performing a family of multiple statistical tests. If this doesn't make any sense to you, please don't use this page! A good place to start reading is the FDR home page.

Please do not use this resource for data to be published without carefully checking the methods and results. As always, our motto is "All care, no responsibility". This software is made available without any kind of guarantee of fitness for any purpose whatsoever. It may NOT work correctly. If you miss the Nobel prize because of my mistake, tough. On the other hand, if you win one, an acknowledgement during your acceptance speech would be appreciated. If you detect reproducible errors, please let know urgently.


Walkthrough Example
I. False Discovery Rate

The first item of importance for this tool is the Desired False Discovery Rate. Please enter the value you would like to use here: .


II. Experimental P Values

This test operates on a set of P values from a family of related tests. Please paste your values in the box below. The tool should be able to recognize values with a variety of seperators, including whitespace (returns, spaces, tabs), and commas. The input format is fairly loose - a sequence of P values separated by some whitespace or a comma. The parser is very dumb, but you should be able to cut and paste from somewhere without too much difficulty. Samples which are known to work include the test data shown in the examples below. Try to cut/paste one of those lists of p values to see what works. Suggestions for better ways to do this are welcomed.


III. To Download or not to download?

Finally, you can configure how the tool communicates the resulting information to your browser. The server can either display the information directly in the browser window, or the results can be downloaded to a location that you specify. Check this box if you would like to be prompted to save the results. The default is to display the information directly in the browser window.

IV. Form submission

When you have completed the above steps, you are ready to submit your data and recive the results. Click the button.


Additional Examples
black_white_1.jpg Data is taken from Table 6 of Curran-Everett, Am J Physiol Reg Int Comp Physiol; 279:R1-R8,2000
FDR
P Values
Download

black_white_1.jpg Data is taken from p6 of a manuscript in preparation from benjamini, kreiger, yekutieli. This time, we will use newlines to seperate the p-values.
FDR
P Values
Download

black_white_1.jpg Data is taken from Table 3 of sabatti,sevice&freimer. We complete our examples by showing the use of commas to delimit the P values.
FDR
P Values
Download