The 2016-2017 IDSC Fellows and their mentors gathered to hear the final presentations on their chosen projects. Presenting first was Michael Durante with his project on “Epigenomic Profiling of Uveal Melanoma”.
Uveal melanoma (UM) is a highly aggressive eye cancer that leads to metastatic death in up to half of patients. Despite great progress in the diagnosis and treatment of UM over recent decades, there has been no corresponding improvement in survival. Michael’s project “Epigenomic Profiling of Uveal Melanoma” focuses on using next-generation sequencing techniques to understand epigenetic regulation of UM tumorigenesis. Using techniques that interrogate the histone modifications and chromatin accessibility, Michael was able to study how the driver mutations change the epigenetic landscape of uveal melanoma. The Pegasus supercomputer was necessary for the analysis of this data, which required advanced mathematical modeling. The guidance he received from his IDSC mentor was instrumental to his success in this project, including access to important resources and experienced faculty that helped shape his PhD thesis work.
His advisors are: J. William Harbour, MD | Bascom Palmer Eye Institute and Stephan Schürer, PhD | CCS Drug Discovery Program Director
Michael Fernandez, formerly of United States Air Force (USAF) Intelligence, is working with the Aerodynamics and CFD Lab on co-flow jet airfoil vortex visualization. Current images are 2D (example at right). Michael’s project is to create 3D visualizations of airfoils and turbomachinery using MATLAB.
Adapting a shallow water equation solver for high-performance computing
Anchen’s project revolves around the numerical solution of the shallow water equations on high performance computers. He will focus on identifying the bottlenecks in the code’s performance, whether it in CPU-bound or memory-bandwidth bound and suggest improvements.
Evolution of the Infrastructural Power of the State: Magdalena Medio, 1982-2002.
Nicolas employed GIS and programming tools to (re)code information held in printed maps or raster images back into a georeferenced relational database. The goal was to profit from cartographic sources, often neglected by researchers, that hold information that might otherwise be lost digitally (i.e., old data that was kept in obsolete magnetic media) but that survives in print or image, often in choropleth or monochromatic patterned maps. Thus, the project does something like a reverse GIS, employing computational tools to navigate sheets of paper and (semi)automatically extract information from pixels, polygons, and patterns representing roads, municipalities, intensities of protests, distribution of public goods, among others.
The project was designed to help him built datasets he employed in his Political Science dissertation about the relationship between State Power and Political Violence in the Magdalena Medio (Colombia). As far as they could, he and his tutor employed open source tools, especially R and Quantum GIS.