Supervised projects

We regularly supervise statistics and mathematical students in writing their BA and MA theses about topics we encounter in our everyday work, we offer internships and supervise PhD theses. Find below a selection of topics.

• Model selection in the framework of multi-state models
• Identification of predictive factors for treatment efficacy in randomized and non-randomized settings
• Strategies for cancer clinical trials with multiple biomarkers
• Extending the multistate framework to incorporate high-dimensional genomic information
• Estimation of effective concentrations from in vitro dose-response data using
• the log-logistic model

• Explorative Identifikation von prädiktiven Biomarkern für Überlebenszeitendpunkte (Exploratory identification of predictive biomarkers for time-to-event endpoints)
• On multiple comparison adjustments in Bayesian clinical trial design
• Risks and benefits of autologous stem cell transplantations in treating elderly patients with multiple myeloma: Competing risks analyses
• Regression models for bounded responses with application to DNA methylation
• Estimating classification probabilities for personalized diagnostics using high-dimensional methylation data
• Comparing clustering methods for Infinium DNA methylation data
• Identification of predictive biomarker signatures based on high-dimensional genomic data
• Evaluation of different pre-screening procedures to reduce dimensionality in high-dimensional Cox models
• Untersuchungen zu Fallzahl und Power für onkologische Studien mit Langzeitüberlebenden (Investigations on sample size and power for oncology studies with long-term survivors)
• A Bias-corrected Estimator for Distance Correlation and its Application in Biostatistics
• Comparing Cox-model based survival tests on large genomic data
• Evaluating Two-Level Data: Comparison of Linear Mixed Model Analysis and ANOVA based on Averaged Replicate Measurements

• Cutpointanalysen für prognostische Faktoren zur Vorhersage komplexer Ereigniszeiten (Cut-point analyses for prognostic factors with respect to complex time-to-event endpoints)
• Multiple imputation under not-at-random missingness with application in prediction with qPCR data
• Konfidenzintervalle für das Hazard Ratio und ihre Verträglichkeit mit dem Log-Rank Test: Eine vergleichende Simulationsstudie (Confidence intervals for the hazard ratio and its consistency with the log-rank test: a comparing simulation study)

• Multiple imputation using Substantive Model Compatible Fully Conditional Specification with time-varying covariates
• Multiple imputation of covariates in Cox proportional hazards models with interaction terms
• Exact unconditional testing for the difference of partially paired proportions
• Performance of different imputation approaches in Label-Free Mass-Spectrometry Proteomics data
• Genome-wide significance and empirical type I error rate of different survival tests - A simulation study
• Implementation of a software pipeline for the analysis of dose response experiments for combination treatments in ray designs.