SAND: automated time-domain modeling of NMR spectra applied to metabolite quantification

Published

January 25, 2024

Author(s)

Yue Wu, Omid Sanati, Mario Uchimiya, Krish Krisnamurthy, Jonathan Wedell, Jeffrey C. Hoch, Arthur S. Edison, Frank Delaglio

Abstract

Developments in untargeted nuclear magnetic resonance (NMR) metabolomics enable the profiling of thousands of bio-logical samples. Exploitation of this rich source of information requires detailed quantification of spectral features. How-ever, the development of a consistent and automatic workflow has been challenging because of extensive signal overlap. To address this challenge, we introduce the soft-ware SAND (Spectral Automated NMR Decomposition). SAND follows upon the previous success of time-domain modeling and automatically quantifies entire spectra without manual interac-tion. The SAND approach uses hybrid optimization with Markov chain Monte Carlo methods, employing sub-sampling in both time and frequency domains. In particular, SAND randomly divides the time-domain data into training and valida-tion sets to help avoid overfitting. We demonstrate the accuracy of SAND, which provides a correlation of ≈0.9 with ground truth on cases including highly overlapped simulated datasets, a two-compound mixture, and a urine sample spiked with different amounts of a four-compound mixture. We further demonstrate automated annotation using correla-tion networks derived from SAND decomposed peaks, and on average 74 % of peaks for each compound can be recovered in single clusters. SAND is available in NMRbox, the cloud computing environment for NMR software, hosted by the Net-work for Advanced NMR (NAN). Since the SAND method uses time-domain sub-sampling (i.e. random subset of time-domain points), it has the potential to be extended to higher dimensionality and non-uniformly sampled data.

Citation

Nature Methods

Volume

96

Issue

5

Pub Type

Journals

Download Paper

https://doi.org/10.1021/acs.analchem.3c03078

Local Download

Keywords

NMR, metabolomics, time-domain analysis, spectral analysis, spectral quantification

Citation

Wu, Y. , Sanati, O. , Uchimiya, M. , Krisnamurthy, K. , Wedell, J. , Hoch, J. , Edison, A. and Delaglio, F. (2024), SAND: automated time-domain modeling of NMR spectra applied to metabolite quantification, Nature Methods, [online], https://doi.org/10.1021/acs.analchem.3c03078, https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=956289 (Accessed September 26, 2024)