Changes in version 1.0.7 (2024-07-03)                  

  - replacing defunct vegan::adonis by vegan::adonis2

                 Changes in version 1.0.6 (2024-01-22)                  

  - Removing unused argument gc_peak_df from align_peaks

                 Changes in version 1.0.5 (2023-01-26)                  

  - Bugfix in choose_optimal_reference that always selected the first
    sample as a reference. Thanks to Heberto del Rio who pointed this
    out on https://github.com/mottensmann/GCalignR/issues/27

                       Changes in version 1.0.3.9                       

  - Speedboost when setting max_diff_peak2mean = 0: In this special case
    there is no need to use a time-consuming iterative approach but
    peaks can be sorted simply based on absolute values. This is
    implemented in two steps. (1) Across all samples, unique retention
    times are extracted, sorted in increasing temporal order and written
    to a template data frame. (2) For each sample, peaks are matched to
    the corresponding row of the template data frame.
  - Small bug fixed that caused problems when plotting x-axis labels in
    gc_heatmap.
  - Added a test for detecting inconsistently ordered retention times
    within samples. Retention times are expected in increasing order,
    starting with the lowest number. If this assumption is violated,
    retention times are reordered and a warning is shown.

                 Changes in version 1.0.3 (2020-08-26)                  

  - Added fill = TRUE as a parameter in utils::read.table when reading
    data from text within internal functions. Loading GC data with
    utils::read.table failed in cases of missing values in a column
    (i.e. empty). This is the correct behaviour as missing data should
    always be coded explicitly by ‘NA’
  - Tibbles are now coerced to data frames
  - Added a new boolean parameter remove_empty for the main function
    align_chromatograms. If samples are empty (i.e.. no peak) this
    parameter allows to remove those cases from the dataset to avoid
    problems in post-hoc analyses. By default FALSE, i.e.. all but the
    blank samples are kept.
  - Added a new boolean parameter permute for the functions
    align_chromatograms and align_peaks. This allows to change the
    default behaviour of random permutation of samples during the
    alignment and might be useful if exact replication is needed.

                 Changes in version 1.0.2 (2018-07-16)                  

  - The accompanying manuscript is published
    https://doi.org/10.1371/journal.pone.0198311 and the citation has
    been added
  - The function beta read_empower2 allows to import HPLC data that has
    been generated using the EMPOWER 2 software

                 Changes in version 1.0.1 (2018-01-16)                  

Bugfixes

  - A bugfix was applied for handling multiple blanks correctly.
  - Progressbars are removed in non-interactive R sessions

                 Changes in version 1.0.0 (2017-09-14)                  

New functions implemented

  - choose_optimal_reference offers an automatism to pick suitable
    references.
  - draw_chromatograms allows to represent a peak list in form of
    chromatogram.
  - remove_blanksallows to get rid of peaks that represent contamination
    after aligning a dataset
  - remove_singletons allows to remove single peaks from the dataset
    after aligning
  - merge_redundant_rows allows to merge rows that were not recognised
    as redundant during the alignment by increasing the threshold value
    for the evaluation of similarity

Algorithm

  - Using pbapply, we implemented progress bars to inform the user about
    the progress and the estimated running time of intermediate steps in
    the alignment of peak lists.
  - By implementing more efficient code, we were able to speed up the
    processing, especially picking references is faster by an order of
    magnitude.
  - Retention times are not rounded to two decimals anymore.
    Calculations still capture a precision of two decimals for
    computational reasons.
  - Within the aligned results, retention times correspond to the input
    values. Linear adjustments are only used internally and are
    documented within the Logfile accessible in the output.
  - Reference samples that are used for the coarse alignment of
    retention times can be picked using a novel algorithm that
    determines the average similarity across the dataset.

warning messages

  - Warnings addressing formatting issues are now more explicit and
    partly rephrased to avoid ambiguity.

Plots

  - Added horizontal axis to barplots summarising peak numbers in
    plot.GCalign.
  - Changed to more prominent colours in binary heatmaps with
    gc_heatmap.
  - The function draw_chromatograms was added as another visualisation
    tool.

Vignettes

  - We included a second vignette that explains the algorithm and the
    supported data in detail.

Documentation

  - Helpfiles were rewritten to enhance clarity.