Language Selection

User Menu

Breadcrumb Navigation

Main Navigation


Reliable gene signatures for microarray classification: assessment of stability and performance

Gene expression data enables sample classification based on subsets of genes (gene signatures). However, approaches that simply train classifiers on the basis of gene signatures have shown significant instability, as discussed in Michiels et al. (2005). Expanding upon this work, we consider all possible combinations of gene selection methods with classification methods. These combinations are then evaluated via random sampling. Signatures that prove to be stable over these samplings provide a more consistent insight into the underlying biological processes that distinguish sample groups.

Show all

Service Menu