|
DFG - SPP 1356, Teilprojekt: E-Pluri-Net: Machine Learning and Evolutionary Comparison of Pluripotency Players and Networks
|
|
|
|
|
|
PROJECT: In the first funding period, "E-Pluri-Net: Machine Learning and Evolutionary Comparison of Pluripotency Players and Networks" is the only in-silico project in the DFG Priority Program "Pluripotency and Cellular Reprogramming". Insight into the key genes of pluripotency and their interrelationships as part of a large network of interaction, regulation (and metabolism) will be pursued by computational analyses of multi-species genome, transcriptome (and proteome) data.
|
|
An electronic representation of pluripotency knowledge allows to put data into a functional context, e.g. to learn biomarkers, to guide experiments, and to identify species-specific pathway evolution. Support vector machines and Bayesian methods will be tailored and extended to learn biomarkers and networks in a (semi-)supervised fashion, given e.g. microarray data as input. Semi-supervised learning allows maximal exploitation of data if labels are scarce. For optimal learning, we will also develop a novel approach that we call bi-classification, that is learning across genes and cell lines at the same time. Inferring the evolution of regulatory regions of genes associated with pluripotency and of networks will contribute to understan!
ding the structural basis of pluripotency. Special emphasis is planned to be given to the analysis of the data generated by the projects of Müller, Schöler, Zenke, Hiiragi, Schroeder and Tanaka. Additional services for the SPP community will include genomic homology search, phylogeny-based protein function prediction and programming/scripting support. The proposed work may also lay the basis for a central data analysis platform, if warranted, for the second SPP funding period.
Projektträger: Deutsche Forschungsgemeinschaft, Bonn
Funding: 144.000 € personnel costs and 9.250 € other costs
Time frame: 2009 - 2011
Contact:
Prof. Dr. Georg Fuellen
Director, Institute for Biostatistics and Informatics in Medicine and Ageing Research -- IBIMA (formerly IMIB -- Institute for Medical Informatics and Biometry) University of Rostock, Medical Faculty Ernst-Heydemann-Str. 8
18057 Rostock, Germany
Phone + 49 381 494-7360
Fax + 49 381 494-7203
http://www.ibima.med.uni-rostock.de/
fuellen [at] alum.mit.edu
|
 This piece of information is also available in German.
|
|
|
Public funded projects 
for projects
uni-rostock.de
