Physics-informed deep learning for plasmonic sensing of nanoscale protein dynamics in solution

Quantifying nanoscale protein secondary structure in aqueous solutions is crucial for understanding protein interactions and dynamics. Deep learning models are adept at predicting protein secondary structures but their ability to model them in aqueous solutions is hindered by data constraints. Here, we present a mid-infrared plasmonic sensor integrated with a synthesized complex-frequency waves (s-CFW) informed convolutional neural network (CNN) to address these limitations. Our sensor enables direct probing of the amide I band in sub-10 nm proteins. By employing s-CFW to amplify target spectral features, the developed physics-informed CNN achieves a mean relative error of less than 0.1 in predicting secondary structure percentages—over twice as accurate as a pristine CNN. Our method enables in-situ and real-time quantification of subtle conformational changes during protein assembly, thereby addressing the issue of data scarcity that currently hinders the development of advanced deep l..., , # Data from: Physics-informed deep learning for plasmonic sensing of nanoscale protein dynamics in solution

Dataset DOI: [10.5061/dryad.qjq2bvqtd](10.5061/dryad.qjq2bvqtd)

Open Source datasets for the manuscript titled \"Physics-informed Deep Learning for Plasmonic Sensing of Nanoscale Protein Dynamics in Solution\".

This data consists of the raw data of the images and the code involved in the article. It contains raw data from nano-fabriction experiments to simulations, and the CNN models.

## Description of the data and file structure

We describe our raw data in the order in which each image appears in the main text. The final dataset includes:

1. Experimental spectra (results of infrared spectrum measurement of SNFs)
2. Simulated spectra (simulated data of proteins under different parameters)
3. Spectra after data augmentation (with noise addition and s-CFW processing)
4. Data generated by the CNN model (prediction results and loss values during the training process)
5. Device and..., We state that our received explicit consent from our participants to publish the de-identified data in the public domain.