Novel Use Of Neural Nets To Convert PSNs To 3D Atomistic Structures


A recent collaboration between Dr. Vy Duong, Dr. Gianmarc Grazioli, and current NCASD member Liz Diessner, published in Biomolecules, illustrates the utility of Neural Networks for retrieving atomistic detail from Protein Structure Networks (PSNs) to reconstruct 3D molecular models. The ability to convert between coarse-grained PSNs and their detailed atomistic structure analogs allows researchers to take advantage of the scalability of network representations to minimize computational constraints on research, without forfeiting useful structural information.