|
Abstract:
|
Affymetrix GeneChip® arrays are widely used as a powerful tool to measure expression levels of thousands of genes simultaneously, particularly to measure relative expression levels between different experiment conditions. They usually require preprocesses of probe level data before analysis at gene level. These preprocesses include background correction, normalization and data transformation, etc. Their objectives are to remove systematic biases and reduce noises in the data. However, it may not always be possible to achieve the two goals at the same time. In this presentation, we demonstrate the trade-off between these two goals by applying different levels of biases in the data preprocess. We find we can significantly reduce noises in the data by including certain levels of biases, and this also improves the sensitivity and specificity of detecting the differentiate gene expression.
|
