Statistical Modelling for Data from Experiments with Short Hairpin RNAs

This paper delivers an example of applying intelligent data analysis to biological data where the success of the project was only possible due to joint efforts of the experts from biology, medicine and data analysis. The initial and seemingly obvious approach for the analysis of the data yielded results that did not look plausible to the biologists and medical doctors. Only a better understanding of the experimental setting and the data generating process enabled us to develop a more suitable model for the underlying experiments and to provide results that are coherent with what could be expected from our knowledge and experience. The data analysis problem we discuss here is the identification of significant changes in experiments with short hairpin RNA. A simple Monte Carlo test yielded incoherent results and it turned out that the assumptions on the underlying experiments were not justified. With a Bayesian approach incorporating necessary prior knowledge from the biologists, we c...

• Type: Conference
• Year: 2010
• Where: IDA
• Authors: Frank Klawonn, Torsten Wüstefeld, Lars Zender
• Keywords: Data Analysis, Data Analysis Problem, IDA 2010, Information Management, Intelligent Data Analysis