Automated descriptors for high-throughput screening of peptide self-assembly

Abstract

We present five automated descriptors: Aggregate Detection Index (ADI); Sheet Formation Index (SFI); Vesicle Formation Index (VFI); Tube Formation Index (TFI); and Fibre Formation Index (FFI), that have been designed for analysing peptide self-assembly in molecular dynamics simulations. These descriptors, implemented as Python modules, enhance analytical precision and enable the development of screening methods tailored to specific structural targets rather than general aggregation. Initially tested on the FF dipeptide, the descriptors were validated using a comprehensive dipeptide dataset. This approach facilitates the identification of promising self-assembling moieties with nanoscale properties directly linked to macroscale functions, such as hydrogel formation.