batchPred: Predict relevant covariates for batch effect adjustment

In NabilaRahman/batchPred: batchPred - Objective prediction of relevant set of batch effects variables

Predict relevant covariates for batch effect adjustment

Description

batchPred predicts relevant batch variables for improving the normalisation and the interpretation of heterogenous datasets.

Usage

batchPred(
  input.edata,
  input.covariates.df,
  input.gold.standard,
  threshold = 1e-04,
  cores = 2
)

Arguments

input.edata	matrix of numeric expression data, where rows are genes and columns are samples.
input.covariates.df	the covariate dataframe. Each covariate is a dataframe column that represents known covariate such as batch number, age or hidden covariates such as a principle components.
input.gold.standard	reference data table consisting of known gene associations. The number of true positives and false positives should be approximately equal. gene IDs in input.gold.standard must match gene IDs in input.edata.
threshold	numeric. Minimum AUC score improvement required for addition to the linear design.
cores	integer. Number of cores / threads. Number of threads is the primary bottleneck for computing the initial covariate order.

Value

batchPred returns a table consisting of with the following columns:

• Covariates: covariate, whose adjustment show greatest improvement in AUC.

• LinearModel: Combination of covariates tested.

• AUC: Area under roc curve (AUC) value after batch adjustment with LinearModel.

• AUCvRaw: difference in AUC value of batch corrected and raw dataset.

• covEffectOnAUC: improvement in AUC value compared to previously tested LinearModel.

NabilaRahman/batchPred documentation built on June 19, 2022, 5:35 a.m.