Coarse-Grained Protein Folding via Variational Inference
2019Shell lab, UCSB
Protein backbone models for template-free folding of 200+ residue CG protein domains and self-assembly in amyloidogenic peptides using variational inference techniques.
Developed simplistic coarse-grained models of hydrophilic and hydrophobic poly-amino acids which can be intelligently combined to produce remarkably accurate backbone models for folding short peptide fragments as well as globular protein domains. This was a proof of principle for protein backbone models designed from amino-acid polymers using only native-contact-based sidechain interactions, successfully folding 200+ residue proteins.
- Paper published in J. Chem. Phys. — later selected as an editor’s pick among the 88 most influential articles of 2019
- GitHub repository (partially documented)
- Coarse-grained backbone forcefield parameters in LAMMPS format (archived in Jekyll site)
This work also produced a post-processing utility that re-orders LAMMPS replica-exchange trajectories by temperature.