Science
I was exposed to scientific research and its corresponding culture of curiosity when I was 16. I was invited to apply and participate in NASA DEVELOP (Digital Earth Virtual Environment and Learning Outreach Project). I had the opportunity to contribute to a wind energy feasibility plan that would later be adopted by then New Mexico’s second largest energy cooperative, Kit Carson Co-op. My team would also get to assist firefighters in developing proactive and pre-emptive measures using satellite imagery to predict where wildfires might begin. Since then, I’ve been unwavering in my certainty that academic scientific research is what I want to do with my life.
Over the following 13 years, my trajectory has become clearer. I worked on a team at UCLA’s Institute of Pure and Applied Mathematics in a collaboration with the Aerospace Corporation optimizing connectivity among satellites. This provided me with a serious basis in programming and algorithm development. I would take those skillsets and that experience into my next research opportunity in the Physics department at UNC-Chapel Hill. Under Dr. Amy Oldenburg, I would help with a collaborative investigation into the link between diabetes and macular edema at UNC’s Ophthalmology department by writing a series of algorithms in MATLAB that: segmented the retinal pigment epithelium and nerve fiber layer, identified ROIs, and subjected those ROIs to quantitative tests. The approach I developed yielded sensitivity and specificity over 90% and I would publish my very first paper as a first author in IEEE Transactions on Biomedical Engineering.
Ironically, one of the tools I investigated but did not use, artificial neural networks, prompted an interest in the real thing. Shortly after graduation, I applied to work at UNC’s Neuroscience Center in Dr. Eva Anton’s developmental mouse neurobiology lab as a research specialist. I got a foundation in wet lab bench experience, mouse colony management, confocal and multiphoton microscopy, surgery, and would later be promoted to lab manager. I made novel observations. Using confocal microscopy, I imaged live tissue sections from embryonic transgenic mice brains and watched GFP labeled interneurons climb the fibers of RFP labeled projection neurons during migration in the cortex. I would perform craniotomies and then using multiphoton microscopy, I observed increased dynamics of GFP labeled cilia near blood vessels in juvenile postnatal transgenic mice. I was also key personnel for the lab and would frequently assist other labs with microscopy projects. After 5 years, I took the opportunity to look at neuroscience research on human subjects using fMRI analysis. I wrote a pipeline for UNC’s human neuroimaging faculty using GNU Make, python, and MATLAB that allows for processing of fMRI data from its raw form (DICOM) through preprocessing to first-level (GLM) analysis.
Now at Carnegie Mellon under Dr. Robert F. Murphy’s guidance, I study protein-protein interactions using a machine learning method referred to as “active learning” and totally computer driven lab automated experimentation. For decades, protein-protein interactions have been studied in small, unsystematic investigations. I have the opportunity using predictive modeling and active learning to intelligently sample sparsely from a large experimental space, defined by multiple independent variables, and model across the entire space what genetic perturbations across a variety of cell types are disruptive via subcellular localization with confocal microscopy.
