Neural networks in pulsed dipolar spectroscopy: a practical guide

TL;DR

This paper provides a practical guide on using deep neural networks for processing pulsed dipolar spectroscopy data, highlighting their design, training, and application in extracting distance distributions in structural biology and related fields.

Contribution

It offers insights into training neural networks with simulated data, discusses architecture choices, and presents a practical workflow for PDS data analysis.

Findings

Neural networks excel at extracting distance distributions from PDS data.

Training with simulated databases enhances robustness and accuracy.

The guide includes options for handling different PDS experimental techniques.

Abstract

This is a methodological guide to the use of deep neural networks in the processing of pulsed dipolar spectroscopy (PDS) data encountered in structural biology, organic photovoltaics, photosynthesis research, and other domains featuring long-lived radical pairs and paramagnetic metal ions. PDS uses distance dependence of magnetic dipolar interactions; measuring a single well-defined distance is straightforward, but extracting distance distributions is a hard and mathematically ill-posed problem requiring careful regularisation and background fitting. Neural networks do this exceptionally well, but their "robust black box" reputation hides the complexity of their design and training - particularly when the training dataset is effectively infinite. The objective of this paper is to give insight into training against simulated databases, to discuss network architecture choices, to describe options for handling DEER (double electron-electron resonance) and RIDME (relaxation-induced dipolar modulation enhancement) experiments, and to provide a practical data processing flowchart.