fexact( varargin )

Version 1.12.0.0 (15.8 KB) by Michael Boedigheimer
Hypergeometric cumulative distribution (left, right and two-tailed) including permutation testing.
4.4
(5)
4.8K Downloads
Updated 20 Dec 2014

View License

Matlab now (2014b) has a two-tailed Fisher exact test (fishertest)! However, I still need a version that can be vectorized to perform 1000s of tests. This implementation performs 2.5e6 one or two-tailed tests per second and is ideally suited for genome-wide association studies or other large problems where 100s of samples and up to 100s of 1000s of tests are being performed.
It has been verified correct for samples sizes up to 100 and spot checked beyond that up to 1000s of samples.

Cite As

Michael Boedigheimer (2024). fexact( varargin ) (https://www.mathworks.com/matlabcentral/fileexchange/22550-fexact-varargin), MATLAB Central File Exchange. Retrieved .

MATLAB Release Compatibility
Created with R2013b
Compatible with any release
Platform Compatibility
Windows macOS Linux
Categories
Find more on Genomics and Next Generation Sequencing in Help Center and MATLAB Answers

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Version Published Release Notes
1.12.0.0

Improved speed for large number of observations.
Slightly improved performance for large number of separate tests.
Updated documentation
1.11.0.0

Improved speed for large number of observations.
Slightly improved performance for large number of separate tests.
Updated documentation
1.10.0.0

Chris Rorden added support for single-sided Leibermeister tests, which corrects for uncertaintity in the marginal sums.
Fixed bug in some two-tailed tests when performing tests on a single observation (doesn't affect vectorized tests)
1.8.0.0

changed to allow permutation testing in 2012a. Added a graphic to help explain contingency table
1.6.0.0

fixed bug in detecting error conditions
1.5.0.0

Fixed a reporting error that may affect results when a single test is done. The error that allowed p-values to be larger than 1 to be reported. p-values smaller than one should not be affected.
1.4.0.0

Has been updated to work with a large sample size if performing a single test.

1.3.0.0

Now can handle large number of observations in a single trial with minimal footprint.
If X = [7.9 7.911; 8 8]*1e6
fexact(x(1),sum(x(:)),x(1)+x(3),x(1)+x(2)) returns 0.0498 for a two-sided test;
1.2.0.0

Improved documentation and help. Improved performance for single tests.
1.0.0.0