We are interested in the applications of machine learning and data science in bioinformatics, biochemistry, molecular biology and human physiology.

The research interests of the lab include:

Studying the differences in signaling and metabolic pathways between subtypes of acute myocardial infarction (heart attack) & developing diagnostic methods for differentiating subtype using molecules in circulation. This research involves high resolution mass spectrometry for the detection and quantification of metabolites and proteins in circulation, statistical classification methods, and machine learning techniques for biomarker discovery and validation. The study of metabolism, cell-cell interactions, and signaling pathways are central to this work. This research is conducted in close collaboration with the lab of Dr. Andrew DeFilippis, a clinician-scientist at the Vanderbilt University and a leading expert in myocardial infarction. The ability to determine the subtype of acute myocardial infarction non-invasively, and at the point-of-care, will help ensure that optimal treatment decisions are made for patients suffering from acute myocardial infarction.

The application of Bayesian statistical methodologies in the field of metabolomics. In particular, we are focused on developing hierarchical Bayesian modeling techniques for Stable Isotope Resolved Metabolomics (a metabolomics technique that allows for a time-resolved view of cellular metabolic processes). Another focus is on the use of Bayesian models for integrating prior biochemical knowledge and mass spectrometry data for determining systems-level changes in metabolic processes given mass spectrometry and nuclear magnetic resonance data. This work is vital in helping researchers who use Stable Isotope Resolved Metabolomics better understand the regulation of metabolic pathways that are dysregulated in cancer and in metabolic diseases (e.g. heart disease and diabetes).

Pathway diagram showing genes and metabolites involved in the generation of oxysterols from cholesterol and other sterols. https://www.wikipathways.org/index.php?title=Pathway:WP4545

Bioinformatics and machine learning algorithms and statistical analysis workflow development for high resolution mass spectrometry. The process of taking the tens of thousands of ion features that are detected in high resolution mass spectrometry analysis and making meaningful inference regarding global changes is both challenging and rewarding. The applications of this work span from cancer biology to the study of compounds present in environmental samples. The latter is of particular regional importance in the study of produced water sources.

Extracted ion chromatogram and experimental MS2 spectra compared to a reference spectra for a pharmaceutical compound detected and quantified using liquid chromatography-mass spectrometry.