Paper ID sheet UCL-INMA-2016.10


Independent component analysis to remove batch effects from merged microarray datasets

Emilie Renard, Samuel Branders, P.-A. Absil
Abstract. Merging gene expression datasets is a simple way to increase the number of samples in an analysis. However experimental and data processing conditions, which are proper to each dataset, generally influence the expression values and can hide the biological effect of interest. It is then important to normalize the bigger merged dataset regarding those batch effects, as failing to adjust for them may adversely impact statistical inference. In this context, we propose to use a "spatiotemporal" independent component analysis to model the influence of those unwanted effects and remove them from the data. We show on a real dataset that our method allows to improve this modeling and helps to improve sample classification tasks.
Key words
Accepted for WABI 2016