Transfer Learning for Protein Structure Classification at Low Resolution

TL;DR

This paper demonstrates that deep learning models can accurately classify protein structures from low-resolution data, enabling faster and cheaper structural analysis with potential for functional prediction.

Contribution

It introduces a neural network approach for classifying low-resolution protein structures using high-resolution training data, and explores the impact of input representations on performance.

Findings

Achieves ≥80% accuracy in protein class prediction from low-resolution structures.

Shows side-chain information may not be necessary for fine-grained classification.

Confirms different structure determination methods share a common feature space.

Abstract

Structure determination is key to understanding protein function at a molecular level. Whilst significant advances have been made in predicting structure and function from amino acid sequence, researchers must still rely on expensive, time-consuming analytical methods to visualise detailed protein conformation. In this study, we demonstrate that it is possible to make accurate ($\geq$80%) predictions of protein class and architecture from structures determined at low ($>$3A) resolution, using a deep convolutional neural network trained on high-resolution ($\leq$3A) structures represented as 2D matrices. Thus, we provide proof of concept for high-speed, low-cost protein structure classification at low resolution, and a basis for extension to prediction of function. We investigate the impact of the input representation on classification performance, showing that side-chain information may not be necessary for fine-grained structure predictions. Finally, we confirm that high-resolution, low-resolution and NMR-determined structures inhabit a common feature space, and thus provide a theoretical foundation for boosting with single-image super-resolution.