Changes in version 1.0.1                        

  - Drift correction and correct_batch_serrf() no longer require the
    optional packages mirai and carrier. They run sequentially unless
    parallel workers are set up with mirai::daemons().
  - Plot axes with large values, such as intensities, show compact
    scientific labels with one exponent per axis (e.g. 0.5E6, 1.0E6,
    1.5E6) instead of a superscript exponent on each label, leaving more
    room for the panels.
  - INTEGRATOR: reading mzML files and peak detection are faster. The
    batch column of the sample list and the uniform_width and baseline
    columns of the transition list are optional, and the valley-drop
    baseline is computed correctly when several features share one
    transition. Step 4 (chromatogram PDFs) finds R on the PATH and, on
    Windows, otherwise uses the newest installed R.
  - MRMhub-viz now caches loaded data and draws chromatograms only as
    they scroll into view, resulting in smoother scrolling and faster,
    automatic updates when plot settings are changed. A new status line
    reports progress and errors.

                        Changes in version 1.0.0                        

First stable release of the MRMhub software framework.

Changes

  - New function set_lipid_class() derives lipid classes from lipid
    feature names using the rgoslin package.
  - Calibration curves support 1/sqrt(x) weighting.
  - External calibration works with only one or two calibrator levels
    too.
  - QC metrics report the number of replicates per QC type (new columns
    n_bqc, n_tqc, n_spl).
  - New release v1.0.0 with INTEGRATOR binaries and the QUANT R package,
    with and without a complete demo project.
  - Various bug fixes and improvements in robustness, performance and
    usability.

                        Changes in version 0.9.9                        

Peer-reviewed version of Burla, Teo et al., Nature Metabolism (2026),
doi:10.1038/s42255-026-01629-2. The version initially submitted for
review was 0.9.2.

Changes

  - New isotope interference correction for precursor (MS1) and
    transition-level (MRM) interferences, based on the LICAR method (Gao
    et al., Anal. Chem. 2021).
  - New batch-correction methods (experimental): empirical Bayes ComBat
    (correct_batch_combat(); Johnson et al., Biostatistics 2007) and
    SERRF random-forest normalization (correct_batch_serrf(); Fan et
    al., Anal. Chem. 2019), complementing correct_batch_centering().
  - New import from and export to the community format mzTab-M
    (import_data_mztab(), save_dataset_mztab(); Hoffmann et al., Anal.
    Chem. 2019).
  - New export to Bioconductor SummarizedExperiment (Morgan et al.) and
    lipidr LipidomicsExperiment objects (Mohamed et al., J. Proteome
    Res. 2020) via save_dataset_summarizedexperiment().
  - More consistent plotting, with plot settings such as point size,
    colours, legend placement and dimensions definable globally via
    mrmhub_set_plot_defaults().
  - New save_plot() writes any plot_*() figure to a file at a defined
    physical size and format, including multi-page PDFs.
  - New save_dataset_rds() and read_dataset_rds() save and load a
    complete MRMhubExperiment, making it easy to share and archive
    datasets with all data, metadata and processing status.
  - Improved console output, processing summaries and more actionable
    error messages.
  - Enhanced up-front validation of function arguments.
  - Improved data and metadata import, with more robust sample and
    feature ID matching, deduplication and stronger schema validation.
  - Various bug fixes and improvements in performance, robustness and
    usability, partly based on user feedback.