FibrilGen: A Python Package for Atomistic Modeling of Peptide β‑Sheet Nanostructures

For de novo peptide-based nanomaterials that rely on the rational design of the peptide primary sequence, a systematic approach to computationally model the diverse and complex potential nanostructures formed by self-assembling peptides would be of considerable value. Here, we present FibrilGen, a bespoke Python package capable of building a broad range of cross-β morphologies at the atomic level. FibrilGen takes a set of input geometrical parameters to initialize peptide packing and fibril morphology, followed by a refinement step to produce a compact assembly. Using FibrilGen, one can, for example, generate a variety of assembled cross-β structures as input for molecular simulations; the package also includes features for geometric analysis of fibril nanostructures and their associated trajectories. We demonstrate the utility of the tool by generating cross-β nanostructures of varying morphologies that compare well with self-assembled arrangements determined from cryogenic electron microscopy and solid state nuclear magnetic resonance spectroscopy. These structures also exhibit conformational stability over microsecond molecular dynamics simulations in aqueous solution. We further assess the capability of the modeling/simulation pipeline to filter out nonexperimental β-sheet fibril structures. Thus, the FibrilGen package provides a route to construction of atomistic supramolecular peptide structures of a variety of possible morphologies, for visualization, simulation, and assessment of interactions and stability.