Department of Medical Informatics & Clinical Epidemiology (DMICE) have been awarded two grants through the NIH Big Data to Knowledge (BD2K) initiative, with the goal of improving health research by training students and researchers to better utilize and generate knowledge from large biomedical datasets. One project is to develop a skills course, with the goal of training students in key issues of big data analysis: problem description, data set discovery, preparation and creation of data sets, and implementation of advanced techniques. This course will be offered in a hybrid learning environment, part online and partially through in-person intensive courses. The initial offering of the BD2K Skills Course will be specifically tailored to summer interns from the School of Medicine. In addition to the skills course, our group will develop Open Educational Resources, in the form of online materials focused on 32 topics that will be freely available to educators and students. The development of these materials is currently underway. The ultimate goal is to train data scientists and improve knowledge discovery.
ODG and the OHSU Library hosted several working sessions to address reproducibility and research communication. The goal was to promote interaction between information scientists and research scientists to support enhanced awareness of data stewardship. Two sessions were held including a "Data Management Open House", where aspects of the data-research cycle were addressed, including issues surrounding research reproducibility and scholarly communication of findings. The slides are available here. An interactive data wrangling session was also held, and a paper describing our outcomes is available here.
This project explored options for enhancing search capabilities for an existing biospecimen search application. Text processing tools were used to map anatomy, pathology, and disease concepts from existing terminologies and ontologies to pathology reports that are currently represented in an unstructured natural text form. The concepts identified in the text were also organized in a relational structure to enable taxonomic and parthood based searches. This was a small exploratory project with a goal of integrating these capabilities in an ongoing effort to expand and integrate OHSU's biospecimen databases. This work was funded by OHSU's Medical Research Foundation.