ChloroDBPFinder: Machine Learning-Guided Recognition of Chlorinated Disinfection Byproducts from Nontargeted LC-HRMS Analysis
Bibliographic record
Abstract
Cl), mass defect, and direct molecular formula (MF) prediction are insufficient for accurate recognition of the diverse Cl-DBPs in real environmental samples. This work proposes a novel strategy to recognize Cl-containing chemicals based on machine learning. Our hierarchical machine learning framework has two random forest-based models: the first layer is a binary classifier to recognize Cl-containing chemicals, and the second layer is a multiclass classifier to annotate the number of Cl present. This model was trained using ∼1.4 million distinctive MFs from PubChem. Evaluated on over 14,000 unique MFs from NIST20, this machine learning model achieved 93.3% accuracy in recognizing Cl-containing MFs (Cl-MFs) and 92.9% accuracy in annotating the number of Cl for Cl-MFs. Furthermore, the trained model was integrated into ChloroDBPFinder, a standalone R package for the streamlined processing of LC-HRMS data and annotating both known and unknown Cl-containing compounds. Tested on existing Cl-DBP data sets related to aspartame chlorination in tap water, our ChloroDBPFinder efficiently extracted 159 Cl-containing DBP features and tentatively annotated the structures of 10 Cl-DBPs via molecular networking. In another application of a chlorinated humic substance, ChloroDBPFinder extracted 79 high-quality Cl-DBPs and tentatively annotated six compounds. In summary, our proposed machine learning strategy and the developed ChloroDBPFinder provide an advanced solution to identifying Cl-containing compounds in nontargeted analysis of water samples. It is freely available on GitHub (https://github.com/HuanLab/ChloroDBPFinder).
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 | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| 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 | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.002 |
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".