News & Events

Professors Kim and Sun receive UNCG FY24 Internal Research Awards

“Machine-learning methods for automated analysis of phenotypic plasticity in mammals”
Assistant Professor Minjeong Kim received a collaborative research award of $25,000 with Assistant Professor Bryan McLean in the Biology department. Phenotypic traits (the morphological, physiological, and behavioral properties of organisms) mediate adaptation and the ability to respond to environmental change over long timescales. Conversely, phenotypic plasticity – the ability of an organism to alter its traits in a non-heritable way – facilitates response to environmental stimuli within a single individual’s lifespan. Phenotypic plasticity is important because it aids in response to rapid global changes. To better understand plasticity, however, there is a need to scale up capture of phenotypic traits at the individual level in wild organisms. The purpose of this proposal is to develop and deploy machine learning methods to quantify seasonal shrinkage and regrowth of skeletal tissues in Sorex shrews – a phenomenon involved in maintaining energetic balance despite fluctuating temperatures and food availability. The project will build research capacity for both PIs; create unique, cross-disciplinary training opportunities for diverse graduates and undergraduates; and support collection of preliminary data for multiple federal grant submissions in the future.

“Structured Clinical Evidence Extraction and Synthesis”
Assistant Professor Yingcheng Sun received an individual research award of $5,000. Clinicians currently spend an average of 67.3 weeks in the task of gathering and synthesizing medical evidence from clinical literature to inform guidelines, health policies, and medical decisions. This process is time-consuming and struggles to keep up with the ever-expanding volume of clinical evidence. In response to this challenge, prof. Sun and his team are developing a deep learning-powered pipeline that transforms free-text clinical evidence into a structured representation, capturing patients, interventions, and outcomes. Utilizing the parsed results from the pipeline, they will create a user-friendly search interface. This interface will empower users to retrieve medical evidence, offering a concise summary of current published evidence relevant to their clinical queries. This initiative will contribute to advancing systems for data extraction and machine reading of lengthy clinical articles.