MRMhub Workflow Examples

Reproducible data analysis notebooks demonstrating the MRMhub targeted LC-MS data-processing pipeline on real datasets. Each workflow is a Quarto notebook combining narrative, executable R and outputs, provided as Supplementary Materials for the MRMhub manuscript (Burla et al., 10.64898/2025.12.20.695370).

Rendered site: slinghub.github.io/MRMhub-workflows.

Workflows

• Dataset 1 — targeted plasma lipidomics (Tan et al. 2022); 503 features across 937 samples.
• Dataset 3 — targeted lipidomics (Chen et al. 2025); 829 features across 4,591 samples.
• Dataset 4 — fully quantitative steroid assay (Panel 1, 15 analytes) with external calibration, low/high QC and EQA samples. MRMhub-QUANT postprocessing: ISTD normalization, calibration, quantification and QC bias/variability against reference targets.
• Dataset 4 — Comparison — the same steroid assay, validating MRMhub against Agilent MassHunter across the INTEGRATOR (peak areas) and QUANT (concentrations) modules.

Data and how to run

The datasets and the full code used to generate this resource are deposited together in the Zenodo record 10.5281/zenodo.15370293 (MRMhub-workflows). This GitHub repository does not contain the Dataset 1 and 3 data due to their sizes — only the Zenodo bundle is complete.

mrmhub-workflows.zip — this complete repository including all data under data/. Unzip it, open the project and run quarto render; nothing else is needed. This is the recommended way to reproduce the workflows.

unzip mrmhub-workflows.zip && cd mrmhub-workflows
quarto render

MRMhub-Dataset1.zip, MRMhub-Dataset3.zip, MRMhub-Dataset4.zip — the raw mzML files together with the MRMhub-INTEGRATOR application and all of its input files (param.txt, feature/transition table, sample list), so peak integration can be re-run exactly as published or with new parameters. These are not needed to reproduce the notebooks: the workflows here start downstream of INTEGRATOR, from the long.csv it produces, which is already included under data/. The record also holds MRMhub-Dataset2.zip, a further dataset not used by these workflows.

Running from a git clone

Without the Datasets 1 and 3 data, a clone renders Dataset 4 only:

git clone https://github.com/SLINGhub/MRMhub-workflows
cd MRMhub-workflows
quarto render Dataset4.qmd            # works; data/dataset-4/ is tracked
quarto render Dataset4_Comparison.qmd # works

For Datasets 1 and 3 (and the manuscript figure, which is built from Dataset 1), use the Zenodo bundle above. Each data/dataset-<n>/ folder carries a README.txt describing its files.

Software environment

The rendered site was produced with mrmhub 0.9.9 (the QUANT module) on R 4.5 and Quarto.

The notebooks render to HTML with no further setup. The PDF versions are typeset with LuaLaTeX, which needs a TeX distribution installed — quarto install tinytex provides a minimal one. Without it, render HTML only, e.g. quarto render Dataset1.qmd --to html.

if (!require("pak")) install.packages("pak")
pak::pak("SLINGhub/MRMhub")

Reproducibility note. To install a fixed version, install from a specific commit or release tag once a version carrying this API is published, e.g. pak::pak("SLINGhub/MRMhub@<tag-or-sha>").

Licence

Code is MIT-licensed; the datasets under data/ and the rendered site in docs/ are CC BY 4.0. See LICENSE.

Citation

Cite the Zenodo deposit, 10.5281/zenodo.15370293 — see CITATION.cff — together with the MRMhub preprint, Burla et al., bioRxiv 2025, 10.64898/2025.12.20.695370.

Contact

Bo Burla (bo.burla@nus.edu.sg) and Hyungwon Choi (hyung_won_choi@nus.edu.sg)