Biophysical Society Thematic Meeting| Santa Cruz 2018

Genome Biophysics: Integrating Genomics and Biophysics to Understand Structural and Functional Aspects of Genomes

Poster Abstracts

12-POS Board 12 Lauren Porter , Loren L. Looger. HHMI, Sterling, USA.

As the number of sequenced genomes increases, the need for computational techniques that identify features of those genomes increases also. Currently, we are developing a computational technique for identifying proteins that switch folds, i.e. remodel their secondary structures in response to cellular changes. Previously, we identified 96 instances of biologically relevant fold switching. These instances occur in response to physiological changes, increase the functionality of the proteome, and occur in all kingdoms of life. Additional evidence suggests that fold- switching proteins are likely to occur more frequently in nature than indicated by the current population of structures in the Protein Data Bank (PDB). Together, these results indicate that protein fold switching is likely more common than currently believed, suggesting that many important functions of proteins remain unknown. Thus we are developing a computational technique that identifies more fold-switching proteins. To do this, we identified two of their characteristic features: incorrect secondary structure predictions and likely independent folding cooperativity. Using these features, our technique correctly identified the fold-switching regions of 14/16 proteins with experimental evidence for fold switching (p < 7.5*10-6, Fisher’s exact test). Additionally, it identified 64/93 proteins (p < 4.7*10-13, Fisher’s exact test) that were likely to switch folds based on homology. Out of the 29 false negatives, 22 result from secondary structure predictions that seem to be biased by overpopulation of common structure in the PDB. We are currently working to decrease these false negatives and winnow down false positives. These improvements—combined with experimental validation of our technique—could lead to the rapid identification of proteins with heterogeneous structural and functional ensembles, which has implications for genomics and human health.

46