The human genome is about 2 meters long and tightly folded into the cell nucleus, a sphere that is about 4 million times smaller than a pinhead. How do cells avoid entangling the DNA and ensure regulation of necessary parts? Biologists study DNA folding through the detection of pairwise physical interactions along the DNA, which results in a 3-by-3 million pixel matrix. Visualized as a heatmap, thousands of local visual patterns become apparent. Studying these patterns is like trying to understand the average layout of parks while viewing countries on a world map. High uncertainty of automatically-derived annotations requires manual inspection of such visual patterns for quality control and sensemaking of underlying biological features. In this talk I will present 3 interactive tools that support different aspects of exploring such datasets at scale: (1) seamless browsing and interactive view composition with HiGlass; (2) local pattern exploration through visual decomposition in HiPiler; and (3) guided navigation using Scalable Insets. I will conclude with a discussion about the generalizability of the tools' underlying concepts and show work in progress on interactive concept learning for epigenomic patterns.

Bio: Fritz Lekschas is a third-year computer science Ph.D. candidate at the Harvard John A. Paulson School of Engineering and Applied Sciences who is developing visual interfaces for scalable exploration of biomedical data in the lab of Hanspeter Pfister, An Wang Professor of Computer Science. Prior to his doctorate program, he visited Harvard Medical School to work with Nils Gehlenborg, Assistant Professor of Biomedical Informatics, on ontology-guided exploration of biomedical data repositories. During his undergrad and master in bioinformatics at the Free University of Berlin, Fritz worked as a research assistant under the supervision of Andreas Kurtz at the Charité – Berlin University of Medicine on data exploration of cell-centric data.