cls.Stats Namespace Reference

Collection of statistics tools. More...

Functions
def	ComputePValueFromSets (setA, setB, total, alternative=TEST_GREATER)
	Compute p-value based on the overlap of two sets and the total background. More...
def	ComputePValueFromMatrix (a, b, c, d, alternative=TEST_GREATER)
	Compute p-value using Fisher's exact test on a 2x2 contingency table: More...
def	PAdjust (p, method)
	Adjust p-values for multiple testing. More...

Variables
string	TEST_GREATER = "greater";
string	TEST_LESS = "less";
string	TEST_2SIDED = "two-sided";
string	MTEST_BONFERRONI = "bonferroni";
string	MTEST_BH = "benjamini-hochberg";

Detailed Description

Collection of statistics tools.

Provides some basic functions used in hypothesis testing.

Authors

Chris X. Weichenberger

Hagen Blankenburg

Date

2014-11-24

Copyright

See LICENSE file.

Function Documentation

ComputePValueFromMatrix()

def cls.Stats.ComputePValueFromMatrix	(		a,
			b,
			c,
			d,
			alternative = TEST_GREATER
	)

Compute p-value using Fisher's exact test on a 2x2 contingency table:

| SetA | Not SetA | Sum

#--------—+------------—+-------------------—+----------—

Set B | a (=SetA&SetB)| b (=SetB\SetA) | SetB

Not Set B | c (=SetA\SetB)| d (total\SetA\SetB) | total\SetB

#--------—+------------—+-------------------—+-----------—

Sum | SetA | total\SetA | total

Parameters

a	Upper left element of the matrix
b	Upper right element of the matrix
c	Lower left element of the matrix
d	Lower right elelement of the matrix
alternative	One of TEST_GREATER, TEST_LESS, TEST_2SIDED.

Returns

A tuple containing the odds-ratio and p-value. The OR may be any positive number, but can also be nan or inf.

ComputePValueFromSets()

def cls.Stats.ComputePValueFromSets	(		setA,
			setB,
			total,
			alternative = TEST_GREATER
	)

Compute p-value based on the overlap of two sets and the total background.

| SetA | Not SetA | Sum

#--------—+------------—+-------------------—+----------—

Set B | a (=SetA&SetB)| b (=SetB\SetA) | SetB

Not Set B | c (=SetA\SetB)| d (total\SetA\SetB) | total\SetB

#--------—+------------—+-------------------—+-----------—

Sum | SetA | total\SetA | total

Parameters

setA	The first set (i.e. positives on one condition)
setB	The second set (i.e. positives on the other condition)
total	The total background set, includes the negatives of both conditions.
alternative	One of TEST_GREATER, TEST_LESS, TEST_2SIDED.

Returns

A tuple containing the odds-ratio and p-value

PAdjust()

def cls.Stats.PAdjust	(		p,
			method
	)

Adjust p-values for multiple testing.

This code is based on the source of the function @c p.adjust in the
statistcal programming language R.
@param p  (@c List) List of float values ideally between 0 and 1. Values
 greater than 1 are truncated to 1 (except @c inf values which are
 kept). Negative values are not truncated and receive the highest weights,
 as they are the lowest 'p'-values. @c nan entries in the list are ignored,
 and they do not count for the number of p-values to correct for. A
 probability of 0 remains 0.
@param method Constants @c MTEST_BONFERRONI or @c MTEST_BH for multiple
 test correction according to Bonferroni and Benjamini-Hochberg,
 respectively.
@return (@c List) List of corrected p-values in the same order as the
 original list and with the same number of entries, including @c inf and
 @c nan entries.

Variable Documentation

MTEST_BH

string cls.Stats.MTEST_BH = "benjamini-hochberg";

MTEST_BONFERRONI

string cls.Stats.MTEST_BONFERRONI = "bonferroni";

TEST_2SIDED

string cls.Stats.TEST_2SIDED = "two-sided";

TEST_GREATER

string cls.Stats.TEST_GREATER = "greater";

TEST_LESS

string cls.Stats.TEST_LESS = "less";