Regensburg 2007 – scientific programme
Parts | Days | Selection | Search | Downloads | Help
BP: Fachverband Biologische Physik
BP 21: High-Throughput Data and their Analysis
BP 21.3: Talk
Wednesday, March 28, 2007, 18:45–19:00, H43
Decomposing gene expression profiles using sparseness and nonnegativity via genetic optimization — Kurt Stadlthanner1, •Elmar Lang1, Ana-Maria Tomé2, Carlos Puntonet3, and Fabian Theis4 — 1Institute of Biophysics, University of Regensburg, Germany — 2DETI/IEETA, Universidade de Aveiro, Portugal — 3Dep. Arqitectura y Tecnología de Computadores, Universidad de Granada, Spain — 4MPI for Dynamics and Self-Organisation, Göttingen, Germany
Nonnegative matrix factorization (NMF) has proven to be a useful tool for the analysis of nonnegative multivariate data. Gene expression profiles naturally conform to assumptions about data formats raised by NMF. However, its cost function is known to have a rather high indeterminacy concerning the component signals extracted. Hence we consider an extension of the NMF algorithm that provides unique solutions whenever the underlying component signals are sufficiently sparse. However, the resulting fitness function is discontinuous and exhibits many local minima, hence we use a genetic algorithm for its optimization. The algorithm is first applied to toy data in order to investigate its statistical properties. Application to a microarray data set related to Pseudo-Xanthoma Elasticum (PXE) then shows that the proposed algorithm performs superior when compared to standard methods with respect to the estimated PXE-related gene clusters.