Loading
A batch effect is a difference in the data caused by when, where or how samples were processed rather than by biology; if it lines up with the outcome, a model learns the batch instead of the disease.
In molecular biology a batch effect occurs when non-biological factors cause changes in the data, and it leads to wrong conclusions when its causes correlate with the outcome of interest (Wikipedia). Moving between RNA-seq and microarray, or between TCGA and a hospital cohort, is the extreme case. ComBat, from Johnson, Li and Rabinovic, is the standard empirical Bayes correction; per-gene z-scoring, quantile normalisation and rank transforms are the alternatives that need no batch labels. Any correction fitted on test data leaks.
Shares Microarray expression data, TPM, FPKM and raw counts (expression units), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Domain shift and domain adaptation (cell line to patient), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Microarray expression data, Domain shift and domain adaptation (cell line to patient), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares TPM, FPKM and raw counts (expression units), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares TPM, FPKM and raw counts (expression units), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Data leakage in model evaluation, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Data leakage in model evaluation, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Data leakage in model evaluation, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.