Tracey Oellerich


Contact

  • E-mail: toelleri@unc.edu.
  • Mail: Department of Mathematics,
    Universirty of North Carolina at Chapel Hill
    120 E Cameron Avenue, CB #3250
    Chapel Hill, NC 27599-3250
  • Office: Mary Ellen Jones, Room 11-.

Curriculum Vitae: [pdf]

About Me

I am a postdoctoral researcher working with Prof. Karin Leiderman at University of North Carolina at Chapel Hill. I completed my PhD in Mathematics at George Mason University under the supervision of Prof. Maria Emelianenko.

I completed my BS in Mathematics and BA in Physics with minors in Statistics and Secondary Education from Wilkes University, Wilkes-Barre, PA in 2016.

I am working on interdisciplinary research at the intersection of Mathematics and Biology, specifically understanding adaptation mechanisms and using tools in data science to infer underlying dynamics and features. I have extended current adaptation criteria to include singular systems previously excluded from consideration. To motivate this extension, I prove several classes of networks emit a singular Jacobian, most notably being networks containing conservation laws. The second half of my work focuses on using techniques in data-science to infer an underlying dynamical system structure and any conservation laws present. I have worked to develop these methods such that they are robust to noise and work in a low-data environment. Using these approaches, the next steps include obtaining experimental data to infer a dynamical system model and associated conservation laws for a protein-protein interaction network. Once the dynamics have been identified, the network can be checked for adaptation.

During my undergraduate years at Wilkes University, I was introduced to the idea of utilizing mathematics in biology applications. I began in the field of computational biophysics where I worked on “Enhanced protein folding through confinement inside a hydrophilic nanopore” under the supervision of Prof. Del Lucent. For this work, we investigated the complexity of protein folding inside a cell, in particular, under confinement. Proteins are large molecules composed of one or more long chains of amino acids. Protein folding is the process in which a protein reaches its functional shape. In this heterogeneous environment, proteins can fold inside of special nanocages called chaperones, which are cone-shaped proteins that assist other proteins in folding or unfolding. Through my work, I have developed a model describing how these cages can accelerate protein folding by modifying the thermodynamics of confined water. I tested these models through distributed computing and molecular dynamics simulation and used state of the art master equation appr oaches to analyze the results and verify my model.

During a study abroad in Padua, Italy, I had the opportunity to participate in the “Bench to Bedside: Translational Molecular Research” course. This course focused on the practical use of genomics, proteomics and bioinformatics for the diagnosis, prevention and treatment of disease. It was co-taught by faculty from US and Italy including experts in the field of medical research on topics such as leukemia, cancer research, infectious diseases, etc. The course utilized hands-on technical workshops and touched on issues related to implementation of new platforms and assays. Through both presentations and experiments, I was given the opportunity to see where the data I work with is collected, the challenges involved in this process, and a different prospective on how it is used in research.

In the summer of 2021, I was given the opportunity to participate as a Data Science intern at the National Center for Advancing Translational Sciences (NCATS) at the National Institutes of Health (NIH). There, I worked on using deep learning to prioritize disease targets for protein-protein interaction networks. We worked on utilizing the data in the Pharos library of known diseases and protein targets to create a ranked list of protein targets by modifying the GuiltyTargets algorithm, a machine learning algorithm that ranks protein targets using gene expression and positive-unlabeled machine learning. This work helps to shave years off the drug development process by eliminating unsuitable targets without the need for testing.

Research Interests

Publications

  1. T. Oellerich and M. Emelianenko. “Towards Robust Data-Driven Automated Recovery of Symbolic Conservation Laws from Limited Data”. (Submitted). arXiv:2403.04889v1
  2. T. Oellerich, M. Emelianenko, M. Pierobon, and E. Baldelli. Learning biological network dynamics from data. (In Preparation).
  3. T. Oellerich, M. Emelianenko, L. A. Liotta, and R. P. Araujo. Biological networks with singular jacobians: their origins and adaptation criteria. Submitted. doi: https://doi.org/10.1101/2021.03.01.433197.
  4. B. Thapa, I. Mazin, P. Suryanarayana, M. Emelianenko, and T. Oellerich. “Devising Momentum-Space Orbital-Free Density Functionals using Machine Learning”. In Preparation.

Research Statement: [pdf]

Research

Research Program Participation

Selected Presentations and Talks

Selected Poster Presentations






Teaching Experience

Honors


Awards

Student Research Talks (StReeTs) at GMU

Society for Industrial and Applied Mathematics - GMU Chapter


Graduate and Professional Student Association (GAPSA) , George Mason University


EXTREEMS-QED Undergraduate Research Program , George Mason University


If you are interested in or need more information about any of the activities listed here, feel free to send me an email. My contact info is listed on the home page.