Automated Micro-Shell Analysis Pipeline for Hyperspectral Nano-IR Data
We developed a fully automated radial micro-shell analysis pipeline for hyperspectral nano-infrared (nano-IR) imaging. Starting with graphene fiber data, we iteratively designed a segmentation and Gaussian-decomposition system capable of tracking peak evolution with sub-micron precision. After validating the pipeline on graphene, we applied it to prion fibrils from Tran et al. (2025), whose Amide-I peaks reflect local secondary structure. Negative and positive radial micro-shells were generated, and each shell’s mean spectrum was fitted using a two-Gaussian model representing β-sheet and α-helix contributions. The extracted parameters (amplitude, center, and width) reveal consistent radial evolution, reflecting structural transitions from the fibril core to the exterior. This work establishes a generalizable framework for automated hyperspectral structural analysis.