With a primary focus on high-dimensional data, I have significant expertise in developing machine learning tools. Much of my work involves constructing Bayesian models, which effectively convert 'prior knowledge', either inherent in the dataset or obtained from external sources, into mathematical terms—more specifically, prior probabilities.
My recent research efforts have centered on the analysis of genetic and epigenetic signals within cell-free DNA assays. This interest in epigenetics led to the development of a pioneering technique known as EPIC-seq, which has broadened our understanding of this complex field.
It's notable that traditional computational methods in cancer genomics often fall short when confronted with an exceedingly low signal-to-noise ratio—a common scenario in cfDNA analyses. As such, there's an emerging need to devise innovative, robust methods capable of overcoming this limitation—a research area that I'm deeply committed to and actively engaged in.