Integrating machine learning with -SAS for enhanced structural analysis in small-angle scattering: applications in biological and artificial macromolecular complexes

Eugen Mircea Anitas

doi:10.1140/epje/s10189-024-00435-6

2023 Impact factor 1.8

Soft Matter and Biological Physics

Eur. Phys. J. E (2024) 47: 39
https://doi.org/10.1140/epje/s10189-024-00435-6

Regular Article - Living Systems

Integrating machine learning with $α$ $\alpha$ -SAS for enhanced structural analysis in small-angle scattering: applications in biological and artificial macromolecular complexes

Eugen Mircea Anitas¹^,2^a

¹ Bogoliubov Laboratory of Theoretical Physics, Joint Institute for Nuclear Research, Joliot-Curie 6, 141980, Dubna, Moscow Region, Russian Federation
² Department of Nuclear Physics, “Horia Hulubei” National R &D Institute for Physics and Nuclear Engineering, Reactorului 30, 077125, Magurele, Ilfov, Romania

^a anitas@theor.jinr.ru

Received: 24 January 2024
Accepted: 16 May 2024
Published online: 3 June 2024

Abstract

Small-Angle Scattering (SAS), encompassing both X-ray (SAXS) and Neutron (SANS) techniques, is a crucial tool for structural analysis at the nanoscale, particularly in the realm of biological macromolecules. This paper explores the intricacies of SAS, emphasizing its application in studying complex biological systems and the challenges associated with sample preparation and data analysis. We highlight the use of neutron-scattering properties of hydrogen isotopes and isotopic labeling in SANS for probing structures within multi-subunit complexes, employing techniques like contrast variation (CV) for detailed structural analysis. However, traditional SAS analysis methods, such as Guinier and Kratky plots, are limited by their partial use of available data and inability to operate without substantial a priori knowledge of the sample’s chemical composition. To overcome these limitations, we introduce a novel approach integrating $α$ $\alpha$ -SAS, a computational method for simulating SANS with CV, with machine learning (ML). This approach enables the accurate prediction of scattering contrast in multicomponent macromolecular complexes, reducing the need for extensive sample preparation and computational resources. $α$ $\alpha$ -SAS, utilizing Monte Carlo methods, generates comprehensive datasets from which structural invariants can be extracted, enhancing our understanding of the macromolecular form factor in dilute systems. The paper demonstrates the effectiveness of this integrated approach through its application to two case studies: Janus particles, an artificial structure with a known SAS intensity and contrast, and a biological system involving RNA polymerase II in complex with Rtt103. These examples illustrate the method’s capability to provide detailed structural insights, showcasing its potential as a powerful tool for advanced SAS analysis in structural biology.

Copyright comment Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

© The Author(s), under exclusive licence to EDP Sciences, SIF and Springer-Verlag GmbH Germany, part of Springer Nature 2024. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.