| Title: | Approximate Unconditional and Permutation Tests |
| Version: | 0.99 |
| Author: | Arjun Sondhi, Ken Rice |
| Maintainer: | Arjun Sondhi <asondhi@uw.edu> |
| Description: | Performs approximate unconditional and permutation testing for 2x2 contingency tables. Motivated by testing for disease association with rare genetic variants in case-control studies. When variants are extremely rare, these tests give better control of Type I error than standard tests. |
| Depends: | R (≥ 3.1.2) |
| License: | GPL-2 |
| Imports: | logistf |
| LazyLoad: | yes |
| Suggests: | knitr, markdown |
| VignetteBuilder: | knitr |
| RoxygenNote: | 5.0.1 |
| Packaged: | 2020-09-03 21:45:22 UTC; asondhi |
| NeedsCompilation: | no |
| Repository: | CRAN |
| Date/Publication: | 2020-09-03 22:12:20 UTC |
Firth AU testing
Description
Calculates approximate unconditional Firth test p-value for testing independence in 2x2 case-control tables. The Firth test requires significantly more computational time than the tests computed in the au.tests function.
Usage
au.firth(m0, m1, r0, r1, lowthresh = 1e-12)
Arguments
m0 |
Number of control subjects |
m1 |
Number of case subjects |
r0 |
Number of control subjects exposed |
r1 |
Number of case subjects exposed |
lowthresh |
A threshold for probabilities below to be considered as zero. Defaults to 1e-12. |
Value
A single AU p-value, computed under the Firth test.
Examples
au.firth(15000, 5000, 1, 0)
Stratified AU testing
Description
Calculates AU p-values for testing independence in 2x2 case-control tables, while adjusting for categorical covariates. Inputs are given as a vector of counts in each strata defined by the covariate(s). Note that computational time can be extremely high.
Usage
au.test.strat(m0list, m1list, r0list, r1list, lowthresh = 1e-12)
Arguments
m0list |
Number of control subjects in each strata |
m1list |
Number of case subjects in each strata |
r0list |
Number of control subjects exposed in each strata |
r1list |
Number of case subjects exposed in each strata |
lowthresh |
A threshold for probabilities below to be considered as zero. Defaults to 1e-12. |
Value
An AU p-value, computed under the likelihood ratio test.
Examples
au.test.strat(c(500, 1250), c(150, 100), c(0, 0), c(10, 5))
AU testing
Description
Calculates approximate unconditional p-values for testing independence in 2x2 case-control tables.
Usage
au.tests(m0, m1, r0, r1, lowthresh = 1e-12)
Arguments
m0 |
Number of control subjects |
m1 |
Number of case subjects |
r0 |
Number of control subjects exposed |
r1 |
Number of case subjects exposed |
lowthresh |
A threshold for probabilities below to be considered as zero. Defaults to 1e-12. |
Value
A vector of AU p-values, computed under score, likelihood ratio, and Wald tests.
Examples
au.tests(15000, 5000, 30, 25)
au.tests(10000, 10000, 30, 25)
Basic testing
Description
Calculates standard p-values for testing independence in 2x2 case-control tables.
Usage
basic.tests(m0, m1, r0, r1)
Arguments
m0 |
Number of control subjects |
m1 |
Number of case subjects |
r0 |
Number of control subjects exposed |
r1 |
Number of case subjects exposed |
Value
A vector of p-values, computed under score, likelihood ratio, Wald, Firth, and Fisher's exact tests.
Examples
basic.tests(15000, 5000, 30, 25)
Stratified permutation testing
Description
Calculates permutation p-values for testing independence in 2x2 case-control tables, while adjusting for categorical covariates. Inputs are given as a vector of counts in each strata defined by the covariate(s). Note that computational time can be extremely high.
Usage
perm.test.strat(m0list, m1list, r0list, r1list)
Arguments
m0list |
Number of control subjects in each strata |
m1list |
Number of case subjects in each strata |
r0list |
Number of control subjects exposed in each strata |
r1list |
Number of case subjects exposed in each strata |
Value
A permutation p-value, computed under the likelihood ratio test.
Examples
perm.test.strat(c(7000, 1000), c(11000, 1000), c(50, 30), c(70, 40))
Permutation testing
Description
Calculates permutation p-values for testing independence in 2x2 case-control tables.
Usage
perm.tests(m0, m1, r0, r1, lowthresh = 1e-12)
Arguments
m0 |
Number of control subjects |
m1 |
Number of case subjects |
r0 |
Number of control subjects exposed |
r1 |
Number of case subjects exposed |
lowthresh |
A threshold for probabilities below to be considered as zero. Defaults to 1e-12. |
Value
A vector of permutation p-values, computed under score, likelihood ratio, Wald, and Firth tests.
Examples
perm.tests(15000, 5000, 30, 25)