PROTEIN STRUCTURE AND DATA SCIENCE
Associate professor, Rutgers University
Modern biology increasingly relies on high-throughput techniques. Yana Bromberg’s work utilizes a variety of computational methodologies, including mining of scientific repositories, machine learning methods, and network analysis, to extract as much useful information from these data as possible. In the genomic sense, this primarily implies correlating phenotypic differences with observed nucleotide sequence variations. On the protein side the challenge is to annotate protein function at reasonable accuracy levels. The whole organism level incorporates all types of evidence to annotate evolutionary history and phenotypes. As an overall goal, the Bromberg group is particularly interested in examining the emergent complexity of proteins in microbial ancestors. They are developing new computational methods for linking protein structure to the evolution of function, particularly to redox functionality of metal binding proteins, with the goal of identifying the oldest protein structures and to time the early evolution of folds. The machine-learning tools built for this purpose will also allow identification of trace-element/metal signatures indicative of the various stages of development of life.