Research Assistant
Raleigh-Durham, North Carolina Area
Developed computational approaches for large-scale genomics data analysis as a part of a cross-disciplinary collaboration project with plant biologists:- Reduced the search space for iron deficiency induced regulators from a set of almost 3000 differentially expressed genes to a testable subset of 7 candidates by applying unsupervised machine learning algorithms to analyze high dimensional whole genome time course datasets- Inferred novel regulatory connections between key stress response genes through the classification and matching of the corresponding temporal expression patterns- Developed a dynamic ODE-based model for simulating gene activity dynamics over time during plant stress response and predicting plant behavior under a variety of conditions- Proposed the types of additional experiments capable of resolving model parameter non-identifiability in an efficient way- Spotted a bias between batches of experimental data, hypothesized the source of the bias based on the technology used, and successfully compensated it with minimal additional experimentation- Mined data from diverse experimental datasets to form prior probability distributions for the model parameters- Quantified uncertainty associated with the model fit- Produced testable predictions for new experiments- Presented research updates at project meetings- Gave 2 guest lectures on current gene expression quantification technologies and strategies for the corresponding data analysis- Leading author of 4 publications (2 published journal papers, 1 journal paper recently submitted for publication, 1 paper in conference proceedings)- Mentored 2 undergraduate students
