Response: Commentary: Data processing thresholds for abundance and sparsity and missed biological insights in an untargeted chemical analysis of blood specimens for exposomics

The commentary by Keski-Rahkonen et al. (1) on our study “Data processing thresholds for abundance and sparsity and missed biological insights in an untargeted chemical analysis of blood specimens for exposomics” (2) needed some additional comparative analyses to clarify the discrepancies noticed for the example peaks that were flagged as missed in the MS-DIAL data processing workflow used in our study. A re-evaluation of the published data matrix available at EBI Metabolights repository accession MTBLS1684 (https://www.ebi.ac.uk/metabolights/MTBLS1684/), from the original study (3, 4) indicated that the LC/MS peaks were reported by identifiers created using neutral masses and retention time (RT) pairs while assuming a proton adduct for all peaks. For example, the peak m/z 412.3035 at RT 5.75 min was reported as “X411.2972.5.749766” and the peak m/z 289.2162 at RT 4.83 min was reported as “X288.2084.4.8316393” in the submitted data matrix. However, in our study (2), MS-DIAL generated data matrix reported the observed m/z values for these peaks. Therefore, those two example peaks were flagged as being missed.

A new comparative analysis has been conducted (https://colab.research.google.com/drive/1eV2ywgLtg0RyJ9qzuVl45KGWlmD0JxPy#scrollTo=c38d63Y8pevi) to expand the discussion. Mass of proton (H+) (1.00784) was added to the reported mass of every peak in the original peak-list to make it comparable to the MS-DIAL generated peak-list. Two new Venn-diagrams (Figure 1) have been created to compare both peak-lists. Among the 623 significant peaks detected using the MS-DIAL peak list, 86% of them were not found among the significant peaks observed from the original peak list (Figure 1B and Supplementary Table S1). Only 7% of all peaks in the MS-DIAL peak list were found in the original peak-list (Figure 1A and Supplementary Table S1). These Venn-diagrams underscore the importance of careful design and review of data processing of untargeted metabolomics datasets from population-scale studies. It also strongly suggests the critical importance of cutoffs for the detection frequency and abundance parameters while generating data matrices from untargeted LC-HRMS datasets. Since MS/MS fragmentation data, cohort data, and the original unfiltered peak lists created without any thresholds for detection frequency and abundances were not made publicly available for the study, additional analyses were not feasible to expand the discussion. The current analysis is limited to unadjusted statistical results. However, the quality of raw LC-HRMS spectral data in the MTBLS1684 study is commendable, this is because of the un-noticeable retention time and signal intensity drifts for a batch of ~ 500 samples. It should be promoted as a benchmarking dataset for teaching and other analyses for comparing performances of data processing software in metabolomics and exposomics.

Figure 1

Statements

Summary

Keywords

metabolomics, exposome, mass spectrometry, birth weight, bioinformatics

Citation

Barupal DK (2022) Response: Commentary: Data processing thresholds for abundance and sparsity and missed biological insights in an untargeted chemical analysis of blood specimens for exposomics. Front. Public Health 10:1003148. doi: 10.3389/fpubh.2022.1003148

Received

25 July 2022

Accepted

28 September 2022

Published

18 October 2022

Volume

10 - 2022

Edited by

Alexandros Siskos, Imperial College London, United Kingdom

Reviewed by

Hector Keun, Imperial College London, United Kingdom

Updates

Check for updates

Copyright

© 2022 Barupal.

This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Correspondence: Dinesh K. Barupal dinesh.kumar@mssm.edu

This article was submitted to Environmental health and Exposome, a section of the journal Frontiers in Public Health

Disclaimer

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.