Machine Learning-Assisted Recognition of Environmental Sulfur-Containing Chemicals in Nontargeted Mass Spectrometry Analysis of Inadequate Mass Resolution
Bibliographic record
Abstract
Sulfur (S)-containing compounds can be unambiguously identified by their distinctive isotope patterns in mass spectrometry (MS) when the instrument has a mass resolution exceeding 500,000. However, many environmental research laboratories that perform nontargeted analysis rely on high-resolution mass spectrometry (HRMS) instruments, such as quadrupole time-of-flight mass spectrometry (QTOF MS). These HRMS instruments typically operate at a mass resolution of less than 50,000. At such limited resolution, confidently recognizing sulfur isotope patterns is challenging. This work develops a machine learning (ML) strategy for recognizing and predicting the number of S present using HRMS at a mass resolution as low as 25,000. We benchmarked our ML strategy on experimental data, where 200 S-containing standard compounds were mixed into complex environmental samples. In positive electrospray ionization (ESI) mode, our ML strategy achieved accuracies ranging from 87.4 to 95.0% for S recognition and accuracies ranging from 86.3 to 96.6% for S number prediction. Notably, the ML method performed similarly well in negative ESI mode. Our ML strategy was further evaluated on an external experimental water dataset where it correctly recognized the presence of S for all 24 previously reported 2-mercaptobenzothiazole disinfection byproducts (DBPs). The developed ML strategy was implemented into SulfurFinder, an R program, to facilitate automated data cleaning, S recognition, and S number prediction in HRMS data. SulfurFinder combined with HPLC-HRMS analysis of a wastewater sample tentatively identified 169 potential S-containing features. Of these, three were confirmed as S-containing pharmaceuticals. An additional S-containing drug was also putatively annotated using molecular networking. The development of SulfurFinder significantly boosts the capability of conventional HRMS to address the challenge of S recognition in the era of exposomics, supporting a wide range of environmental applications.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".