Hi, I'm Kai!
I have advanced training in biostatistics and bioinformatics and my work is fundamentally motivated by applications to real-life genomic research questions through close collaborations involving researchers from a variety of scientific backgrounds. I was a Postdoctoral Fellow under the supervision of Drs. Ingo Ruczinski and Jeff Leek in the Department of Biostatistics at the Johns Hopkins Bloomberg School of Public Health after I had obtained his Ph.D. in Statistics from the University of Dortmund, Germany. During my doctoral studies, I developed new methodological approaches for modeling survival data in the presence of high-dimensional covariates with the goal of improving prediction accuracy and interpretability through the integration of gene expression data and prior biological knowledge on groups of genes. These approaches were subsequently applied to microarray data of individuals with breast cancer to identify gene groups associated with survival times. The results showed improved prediction performance compared to classical models using clinical information or gene expression measurements as covariates.
My current research focuses on developing new statistical methods and software tools for the integrative analysis of high-throughput genomic data, including sequencing and proteomic data (e.g. RNA-sequencing, DNA genotype, and iTRAQ/TMT proteomics data). I develop methods for pre-processing such genomic data, and software to reproducibly execute these protocols. Currently, the joint analysis of RNA-sequencing data and DNA genotypes (delineated by genomic arrays or next generation sequencing) to detect patterns of transcript expression related to specific genetic variants - known as eQTLs, or expression quantitative trait loci - is one of my main research areas. In the future, I will extend these methods to also incorporate methylation and proteomic data in our quest to fully understand the underlying biology.
I am also very interested in helping to bridge the sometimes existing gap between statisticians and scientists, developing easy to use tools based on new or existing sound statistical principles. For example, I have developed open source software for the normalization of isobaric mass labeled proteomic data, with subsequent inference based on moderated test statistics.