DataPrep_test: Preparing the covariates/responses for function...

In terrysun0302/FastMix: An OLS-based approach for fitting LMER with applications to gene expression deconvolution analysis

Description

This function is designed to help prepare the "testing" data set for computing the discriminant scores and help validating a classifier trained by discriminant analysis.

The difference between 'DataPrep()' and 'DataPrep_test()' is that in the latter case, we do not know the true response status of the testing data. However, we need to create a "Response" for these subjects in order to do a proper standardization for the covariates. These two sets of data are standardized according to two different "hypothetical" responses: Response==0 (Data_0), Response==1 (Data_1). Due to the use of standardization in the FastMix model, the processed "Response" in both 'Data_0' and 'Data_1' typically do not equal 0 and 1; instead, they equal to $+a$ and $-a$, where $a$ is a constant for all subjects, respectively.

Usage

  DataPrep_test(GeneExp, CellProp, Demo, train_response, include.demo=TRUE, w="iid")

Arguments

GeneExp	'GeneExp' is a m by n dimensional gene expression matrix, where m is the number of genes, and n is the number of subjects.
CellProp	'CellProp' is a n by K dimensional matrix of cell proportions, where K is the number of cell proportions.
Demo	Demo is a n by p dimensional matrix of clinical and demographic variables to be tested, where p is the number of covariates.
train_response	The column of reponse variable in the training data.
include.demo	Whether the demographical covariates should be included as the main effects in the model or not. Default to TRUE.
w	The weight matrix. The default value of 'w' is 'iid', which refers to a diagonal weighting matrix that is appropriate for i.i.d. data.

Value

Data_0	The data set with pseudo respone Res = 0.
Data_1	The data set with pseudo respone Res = 1.

Note

This function checks the numerical singularity of 'X', the combined covariate matrix. If the smallest singular value is less than '1e-7', it stops and asks the user to reduce the complexity of the model.

Author(s)

Hao Sun

Examples

  ## load the data example
  data(dat_test)
  data(dat_train)

  n2 <- nrow(dat_test$Demo)

  gnames <- rownames(dat_test$GeneExp); m <- nrow(dat_test$GeneExp)
  if (is.null(gnames)) {
    rownames(dat_test$GeneExp) <- gnames <- paste0("Gene", 1:m)
  }


  ## preparing the covariate/response: the i.i.d. case
  test_data = DataPrep_test(dat_test$GeneExp, dat_test$CellProp, Demo=dat_test$Demo[,1], train_response = dat_train$Demo[,2], include.demo=TRUE)

  Data_0 = test_data$Data_0
  Data_1 = test_data$Data_1

  ## an example of the weighted case.
  nsamples <- nrow(dat_train$Demo)
  # randomly generate some weights
  tnmr <- runif(nsamples, 1, 5)     
  w = diag(tnmr)

  ## 
  test_data = DataPrep_test(dat_test$GeneExp, dat_test$CellProp,
  Demo=dat_test$Demo[,1], train_response = dat_train$Demo[,2],
  include.demo=TRUE, w=w)

terrysun0302/FastMix documentation built on Nov. 14, 2019, 4:54 a.m.