DREAM6/FlowCAP challenge 2011

The team "Admire-LVQ" composed of Michael Biehl   (Adjunct Hoogleraar of Computer Science, Johann Bernoulli
Institute) Kerstin Bunte (PhD student, Johann Bernoulli Insitute) Petra Schneider (JBI alumni, now Postdoctoral Researcher at the
University of Birmingham/UK) has achieved best performance in the DREAM6/FlowCAP challenge 2011. This challenge was organized
jointly by the DREAM project ("Dialogue for Reverse Engineering Assessments and Methods") and the FlowCAP project ("Flow
Cytometry: Critical Assessment of Population Identification").

In the Molecular Classification of Acute Myeloid Leukaemia (AML) Challenge, the goal was to identify cases of AML based on flow
cytometry data. Example data of diagnosed patients were provided, the teams had to hand in predictions with respect to a set of 180 patients whose diagnosis was unvailable to the participants.

The team applied a machine learning technique which has been developed within the Intelligent Systems group as part of an NWO supported research project:  "Adaptive Distance Measures in Relevance Learning Vector Quantization" (Admire-LVQ).
The team achieved the best possible performance with 100% correct prediction of all AML cases in the test set.
Michael Biehl was invited to present and discuss these results at the RECOMB::DREAM 2011 conference in Barcelona, October 14th 2011.

For further information, please visit:

www.cs.rug.nl/~biehl
www.the-dream-project.org
http://flowcap.flowsite.org