An article by Christopher Schröder and Sven Rahmann about estimating parameters of beta mixture models, which has applications in determining the methylation state of genomic regions, has been accepted at WABI 2016 and will be presented at the conference in Aarhus (Danmark), August 22-24, 2016. The paper will be available in the WABI 2016 proceedings (LNBI series, Springer Verlag) in August 2016.
A hybrid parameter estimation algorithm for beta mixtures and applications to methylation state classification by Christopher Schröder and Sven Rahmann
Mixtures of beta distributions have previously been shown to be a flexible tool for modeling data with values on the unit interval, such as methylation levels. However, maximum likelihood parameter estimation with beta distributions suffers from problems because of singularities in the log-likelihood function if some observations take the values 0 or 1. While ad-hoc corrections have been proposed to mitigate this problem, we propose a different approach to parameter estimation for beta mixtures where such problems do not arise in the first place. Our algorithm has significant computational advantages over the maximum-likelihood-based EM algorithm. As an application, we demonstrate that methylation state classification is more accurate when using adaptive thresholds from beta mixtures than non-adaptive thresholds on observed methylation levels.
Bioinformatics Analysis of Heterogenous Data Reveals Characteristic Mutational Landscapes of Neuroblastoma Relapses, GCB 2015 in Dortmund
Marc Schulte, Johannes Köster, Daniela Beisser, Corinna Ernst, Christopher Schröder, Alexander Schramm and Sven Rahmann
Neuroblastoma is a malignancy of the developing sympathic nervous system that causes 15% of childhood cancer-related mortality. However, in the vast majority of cases death results not from the initial disease manifestation but rather from metastasis or recurrence.
Systematic search for genomic alterations in primary neuroblastomas has shown low genetic complexity, with significant mutations in only a very few genes. This study explored the genomic landscape of relapsing neuroblastoma in order to evaluate ‘driver’ mutations to be exploited as therapeutic targets.
Imprinting of the human RB1 gene is due to the presence of a differentially methylated CpG island (CGI) in intron 2, which is part of a retrocopy derived from the PPP1R26 gene on chromosome 9. The murine Rb1 gene does not have this retrocopy and is not imprinted. We have investigated whether the RB1/Rb1 locus is unique with respect to these differences.