fastJT.select {fastJT}

Conduct k-Fold Cross-Validation based on Jonckheere-Terpstra Test Statistics

Description

A method to conduct k-fold cross-validation based the Jonckheere-Terpstra test statistics for large numbers of dependent and independent variables, with optional multi-threaded execution. The data are divided into k subsets, one of which is withheld while the remaining k-1 subsets are used to test the features. The process is repeated k times, with each of the subsets withheld exactly once as the validation data. The calculation of the standardized test statistics employs the null variance equation as defined by Hollander and Wolfe (1999, eq. 6.19) to account for ties in the data.

Usage

fastJT.select(Y, X, cvMesh=NULL, kFold=10L, selCrit=NULL, outTopN=15L, numThreads=1L)  	

Arguments

Y

A matrix of continuous values, representing dependent variables, e.g. marker levels or other observed values. Row names should be sample IDs, and column names should be variable names. Required.

X

A matrix of integer values, representing independent variables, e.g. SNP counts or other classification features. Row names should be sample IDs, and column names should be feature IDs. Required.

cvMesh

A user-defined function to specify how to separate the data into training and testing parts. The inputs of this function are a vector representing the sample IDs and kFold, an integer representing the number of folds for cross validation. The output of this function is a list of kFold vectors of sample IDs forming the testing subset for each fold. The default value is NULL, and if no function is specified, the data are partitioned sequentially into kFold equal sized subsets. Optional.

kFold

An integer to indicate the number of folds. Optional. The default value is 10.

selCrit

A user-defined function to specify the criteria for selecting the top features. The inputs of this function are J, the matrix of statistics resulting from fastJT, and P, the matrix of p-values from pvalues(J). The output is a data frame containing the selected feature ID for each trait of interest. Optional. The default value is NULL, and if no function is specified, the features with the largest standardized Jonckheere-Terpstra test statistics are selected.

outTopN

An integer to indicate the number of top hits to be returned when selCrit=NULL. Unused if selCrit!=NULL. Optional. The default value is 15L.

numThreads

A integer to indicate the number of threads to be used in the computation. Optional. The default value is 1L (sequential computation).

Value

Three lists of length kFold

J

A list of matrices of standardized Jonckheere-Terpstra test statistics, one for each cross validation.

Pval

A list of matrices of p-values, one for each cross validation.

XIDs

A list of matrices of the selected feature IDs, one for each cross validation.

Note

Rows (samples) are assumed to be in the same order in X and Y.

References

Hollander, M. and Wolfe, D. A. (1999) Nonparametric Statistical Methods. New York: Wiley, 2nd edition.

See Also

fastJT

Examples

# Generate dummy data	
num_sample  <- 100
num_marker  <- 10
num_feature <- 500
set.seed(12345)
Data <- matrix(rnorm(num_sample*num_marker), num_sample, num_marker)
Features <- matrix(rbinom(num_sample*num_feature, 2, 0.5), num_sample, num_feature)
colnames(Data) <- paste0("Var:",1:num_marker)
colnames(Features) <- paste0("Ftr:",1:num_feature)

res <- fastJT.select(Y=Data, X=Features, cvMesh=NULL, kFold=5, 
                     selCrit=NULL, outTopN=5, numThreads=1)
res