prolfquapp 2.6.1

• Repeated-measures annotation processing now evaluates its subject-column and repeated-design checks as a single scalar condition.
• Shared Quarto reports now include the current FGCZ responsive figure-grid styling and opt-in full-width layout.
• Protein annotations are now joined to peptide and PTM results by protein ID, so every quantified feature retains its protein metadata without losing or multiplying result rows.
• The FGCZ Quarto dependency now follows its canonical fgcz/fgczQuartoTemplate upstream and camel-cased package name, including the latest shared report assets and tab/download controls.
• The experimental-design survey now links to NIST’s maintained guidance on blocking factors instead of an obsolete external URL.
• Package builds now install Quarto visual abstracts as dedicated runtime assets instead of vignette output, eliminating an R CMD check NOTE while preserving runtime-rendered report overviews.
• Quarto reports now vendor the dynamic horizontal FGCZ toolbar: it stays below the visible banner, pins to the top-right while scrolling, and expands its Find/Download text labels on hover or keyboard focus.
• The Docker image now installs the prolfqua Firth-model performance fix, so firth_nested analyses with very large peptide effects complete instead of spending days on unused coefficient profile-likelihood intervals.

prolfquapp 2.6.0

• The full Docker image now includes the fixed-size Quarto report figures and interactive widgets introduced in 2.5.1.

prolfquapp 2.5.1

• Quarto report figures now use fixed display dimensions based on the historic six-inch figures, so QC and differential-expression diagnostics remain a readable size and fit vertically on ultrawide screens.

prolfquapp 2.5.0

• Docker releases now derive their build mode from the semantic version: X.Y.0 tags run the full multi-architecture build and runtime checks, while later X.Y.Z patch tags update the packages declared in Remotes and rebuild prolfquapp on the matching X.Y.0 image. An urgent patch release fails if that full base image is unavailable or if the Docker environment or package dependencies changed since X.Y.0.
• Docker builds now use Posit’s Ubuntu Noble R base image and native Posit Package Manager binaries on AMD64 and ARM64, avoiding lengthy source compilation while retaining Arrow’s zstd support.
• Docker images are now published to the GitHub Container Registry (ghcr.io/prolfqua/prolfquapp) instead of Docker Hub. Images from earlier versions remain available on Docker Hub; new releases are published to GHCR.

prolfquapp 2.4.1

• Quarto analysis reports now render standard static and interactive figures at a centred two-thirds of the content width. Intentional multi-column figure layouts, data tables, and compact Overview visual abstracts are unchanged.

• All five Quarto reports now open with a compact, report-specific visual Overview: three summary cards show the number of samples, experimental groups, and quantified proteins (or the analysed feature type), followed by the visual abstract. The reports finish with a two-subtab Session Info area: Report provenance records the B-Fabric/input context, creator, timestamp, software/model, and package version, while R session info contains only sessionInfo(). The visual abstracts are packaged with the reports, so the same layout is retained in runtime-rendered HTML output.

• Quarto visual abstracts are now copied as individual vignette assets, so make build-vignettes no longer fails when devtools stages report assets into doc/.

• Package builds now exclude Quarto’s transient vignettes/.quarto freeze cache, avoiding nonportable tar-path warnings and cache files in source tarballs.

• Package and vignette builds now synchronize the four FGCZ Quarto assets from fgczquartotemplate, so generated reports use the current shared toolbar and styling.

• Quarto vignette extraction now uses safe defaults for report-local conditional metadata, preventing spurious missing-object errors while creating the companion .R sources.

• Differential Expression Analysis Quality Control report (DiffExpQC_R6_tabset.qmd) now presents its report/analysis metadata (Workunit, Order, Project, generated-by, timestamp, software, model, package version) once — as a table in a final Session Info tab, alongside sessionInfo() — instead of duplicating it in a top-of-page callout.

• Quarto reports now ship the updated right-aligned Find / Download toolbar asset, positioned below the FGCZ banner around one-quarter of the viewport height from the top; plot ZIP downloads can include Order/Workunit metadata and a current timestamp.

• The documentation website now uses Quarto repository source links instead of the broken altdoc code-links: true sidebar entry, so the source link points to GitHub instead of undefined.

• Interactive Plotly subplots now fade non-hovered keyed traces, so abundance-density curves from prolfqua become easier to inspect sample by sample.

• Quarto reports now render reliably under R CMD check and in fresh installs. The reports are rendered via fgczquartotemplate::fgcz_render(), which stages the FGCZ template assets (_metadata.yml, fgcz.scss, fgcz_header_quarto.html, fgcz-plot-finder.html) next to the report from the installed fgczquartotemplate package. Previously the reports used the fgczquartotemplate-html Quarto extension and relied on the _extensions/ tree being shipped into the installed doc/, but the vignettes/.install_extras rule never actually shipped it, so rendering failed when no _extensions/ directory was present (e.g. under R CMD check). The reports are now plain format: html documents styled by a directory-level _metadata.yml, with the Find/Download toolbar wired via include-after-body: fgcz-plot-finder.html; fgczquartotemplate was added to Imports.

• The explanatory info callouts in the Quarto reports (e.g. “Why look at the fold-change and p-value distributions?” in the tabbed DEA report, and the “About this report” note in the QC & sample-size report) are now collapsed by default, so the reports open with a cleaner overview and readers expand a note only when they want it.

• Restructured the “Protein Signal Intensities within Groups” report (QC_ProteinAbundances_tabset.qmd): the Protein abundances tab now leads with the iBAQ signal figure, followed by the table, with the column descriptions and the “What is the iBAQ signal?” explanation moved into collapsed info boxes; Order/Workunit metadata moved from the top of the page into a new final “Session Info” tab (alongside sessionInfo()). Also fixed the column-description list, where the nrMeasured_<GroupName> / meanAbundance_<GroupName> / signal_percent_<GroupName> names had their <GroupName> suffix silently dropped (it was parsed as an HTML tag); the group-suffix placeholders now render.

• Documentation website Articles menu cleaned up: the entries are now ordered Differential Expression Analysis, its Quality Control report, the tabbed Differential Expression report, Protein Signal Intensities within Groups, Quality Control & Sample Size Estimation, and the Auxiliary meeting-agenda article last. The Grp2Analysis_V2_SE_tabset report was retitled from “Differential Abundance Analysis” to “Differential Expression Analysis (Tabbed Report)” so it is clearly the tabbed variant of the main DEA report. The stub prolfquapp.Rmd introduction vignette (an incomplete skeleton duplicating the README) was removed.

• Tabbed DEA report (Grp2Analysis_V2_SE_tabset.qmd): rewrote the cryptic UpSet-plot captions (feature-detection overlap between groups; significant / increased / decreased features shared between contrasts) so each states that it is an UpSet plot and explains what the intersection bars, dot matrix, and set-size bars represent, and clarified the MA-plot and significant-feature heatmap captions.

• The default contrast model is now lm_impute instead of lm_missing. lm_impute refits proteins whose per-protein linear model failed or was singular by imputing at the limit of detection with borrowed variance (flagging rescued rows as lod_imputed), whereas the deprecated lm_missing substituted group means without a model fit. This changes default DEA results for proteins that could not be fit directly, removes the prolfqua deprecation warning emitted on every default run, and applies to make_DEA_config_R6(), run_make_yaml(), the prolfqua_yaml.sh --model default, and the CompoundDiscoverer DEA entry point. lm_missing remains available as an explicit model = / --model choice.

• The documentation website is now built with altdoc (Quarto Website backend) instead of pkgdown. pkgdown’s tweak_tabsets step crashes on the Quarto panel-tabsets used by the tabbed report vignettes; altdoc renders the vignettes natively through Quarto, so the tabset reports appear on the site with their tabs intact. Removed the obsolete make quarto/render-quarto preview targets from the Makefile. # prolfquapp 2.4.0

• The DEA results index.html (write_index_html) is now rendered as an FGCZ Quarto entry page with separate deliverable tables for HTML reports, Excel workbooks, ORA input gene lists, and GSEA rank files, including captions, file sizes, report descriptions, and Excel-content descriptions.

• Report HTML output filenames now match their Quarto source rather than the workunit-based DE_/QC_ scheme: a DEA run writes Grp2Analysis_V2_R6.html, Grp2Analysis_V2_SE_tabset.html, DiffExpQC_R6_tabset.html, and QCandSSE_tabset.html; a QC run writes QC_ProteinAbundances_tabset.html and QCandSSE_tabset.html — all inside the per-workunit Results_WU_<workunit>/ (or QC output) folder. The index page continues to label links with descriptive report titles, not filenames.

• Renamed the Quarto report vignettes to drop the now-redundant _quarto suffix (every report is Quarto): the primary DEA report is Grp2Analysis_V2_R6.qmd, and the tabbed reports carry a _tabset suffix — DiffExpQC_R6_tabset.qmd, QCandSSE_tabset.qmd, and QC_ProteinAbundances_tabset.qmd (Grp2Analysis_V2_SE_tabset.qmd already followed this convention). The report HTML filenames produced by the DEA and QC pipelines are unchanged.

• Reviewed every figure and table caption across the five Quarto reports against the FGCZ searchable-caption rule and rewrote the 16 that were vague or inaccurate into specific scientific labels (naming the measured quantity, what points/bars represent, the axes/encoding, grouping, and transformations). This corrects two DEA figures that were labelled “Venn diagram” but actually draw UpSet plots, disambiguates the two previously identical per-sample protein-count captions (proteins with ≥1 vs ≥2 peptides), and fixes the protein-abundance figure caption (its x-axis is the abundance-rank percentile, not the signal contribution).

• SE tabset DEA report (Grp2Analysis_V2_SE_tabset.qmd): the fold-change / p-value figure now carries a specific, searchable scientific caption (replacing the vague “Fold-change and p-value summaries.”) and is preceded by a callout explaining why both the fold-change distribution and the p-value distribution are inspected as model diagnostics (centred-near-zero fold-changes; approximately uniform p-values under the null).

• Removed the retired R Markdown report sources now that the pipelines render Quarto only: Grp2Analysis_V2_R6.Rmd, DiffExpQC_R6.Rmd, QC_ProteinAbundances.Rmd, and QCandSSE.Rmd. The legacy DEAReportGenerator$render_DEA() method and the render/markdown/markdown_qc arguments of write_DEA_all() (which drove the R Markdown rendering) were removed, and the exported helper copy_DEA_R6_Files() (which copied the R Markdown templates into the run’s input folder) was removed. The two non-report R Markdown vignettes (prolfquapp.Rmd, Auxiliary_ExDesignSurvey.Rmd) are unaffected.

• The DEA and QC command-line pipelines now render Quarto reports only; the R Markdown reports are no longer rendered. The DEA run produces DE_<workunit>.html (the R6 Quarto report, primary), DE_<workunit>_tabset.html (the SummarizedExperiment tabset overview), QC_<workunit>.html (the differential-expression QC report, now a tabbed Quarto report), and QC_sampleSizeEstimation.html (the sample-size estimation report, now also produced during a DEA run). The QC run produces proteinAbundances.html and QC_sampleSizeEstimation.html from their Quarto reports. Report rendering is centralized in one place and each report renders independently, so a single report failure warns without aborting the run or dropping the others. The .Rmd report sources remain in the package but are no longer used by the pipelines.

• New Quarto report DiffExpQC_R6_quarto.qmd: a tabbed port of the differential-expression QC report (DiffExpQC_R6.Rmd), with Missing Values, Variance, and Differential Expression tabs, styled with the FGCZ Quarto extension. It builds from a serialized DEAnalyse .rds (falling back to example_deanalyse()).

• QCandSSE_quarto.qmd: the Sample Size Calculation section now splits into one sub-tab per tested log2 fold-change (0.59, 1, 2), each showing the sample-size bar chart and a smaller per-effect-size table, instead of stacked sub-figures and one wide table.

• QCandSSE_quarto.qmd: when the data contain no missing values (so the missing-value heatmap is empty) the report now shows an explicit “no missing values” placeholder in place of the heatmap, instead of dropping the figure and leaving a broken “Figure ??” cross-reference.

• QC_ProteinAbundances_quarto.qmd is now a two-tab report with no table of contents: the first tab (shown by default) holds the interactive protein table and the protein signal-contribution plot (the crosstalk table stays interactive), and the per-sample protein-count barplot moves to a second tab.

• QC runs now also produce the Quarto sample-size report (QC_sampleSizeEstimation_quarto.html) alongside the existing R Markdown QC_sampleSizeEstimation.html. It is rendered from a serialized copy of the QC data and receives the B-fabric project, order, and workunit identifiers, so its Workunit/Project/Order header is populated (falling back to “n/a” when an identifier is unset). Rendering is skipped with a warning, without failing the QC run, when the Quarto CLI or the installed report sources are unavailable.

• example_deanalyse() now sets example B-fabric identifiers (project/order/workunit), so the differential-expression Quarto report renders with a populated project context instead of blank fields when built from the bundled example.

• QCandSSE_quarto.qmd layout tweaks: the per-sample feature counts and the sample-overlap UpSet plot are now shown side by side, as are the missing-value histogram and the missingness heatmap. The sample-size figure is split into one sub-panel per log2 fold-change (0.59, 1, 2) instead of a single cramped facet grid, giving each effect size full height.

• The QCandSSE_quarto.qmd report is now laid out as a tabbed report (Quarto panel-tabset), matching Grp2Analysis_V2_SE_tabset.qmd. Top-level tabs are Introduction, Quality Control, Sample Size Calculation, Sample Mapping, and Session Info; the Quality Control tab groups Feature Detection, Missing Values, Abundance Distributions, Variance, and Sample Structure into sub-tabs, and the Variance sub-tab splits the raw-intensity coefficient of variation and the transformed-scale standard deviation into their own sub-tabs. All figures and tables keep their cross-reference numbers across tabs.

• New Quarto ports of the QC vignettes, QC_ProteinAbundances_quarto.qmd and QCandSSE_quarto.qmd, styled with the vendored FGCZ Quarto extension (left table of contents, Find/Save toolbar) to match the other Quarto reports. They build from example data by default (example_qc_generator() and the data_ionstar example) and take file parameters (pap_file, qc_data_file) for real inputs. The existing QC_ProteinAbundances.Rmd / QCandSSE.Rmd are unchanged and the QC pipeline still renders them.

• Restructured the QCandSSE_quarto.qmd report for clearer scientific storytelling: consistent section hierarchy (Quality Control → Feature Detection / Missing Values / Abundance Distributions / Variance / Sample Structure, then Sample Size Calculation), an introduction roadmap and a visible caveat callout, every figure and table now referenced in the text, and an explicit coefficient-of-variation → standard-deviation → sample-size narrative. The sample correlation heatmap and the transformed-intensity overview heatmap are now shown side by side. Coefficient-of-variation and standard-deviation summary tables are rounded (2 and 3 decimals), sample-size counts are shown as integers, and a Session Info section with a provenance line was added.

• Both Quarto reports now live in vignettes/ as the single source of truth. The SummarizedExperiment tabbed report (Grp2Analysis_V2_SE_tabset.qmd) moved out of inst/templates/quarto/ into vignettes/ and adopts the vendored FGCZ Quarto extension styling (with the Find/Save toolbar and a left table of contents), matching Grp2Analysis_V2_R6_quarto.qmd. Both reports are shipped into the installed package’s doc/ directory, and the DEA CLI renders them from there: each run now also writes DEAnalyse.rds next to SummarizedExperiment.rds and produces the Grp2 differential-expression Quarto report (*_quarto_dea.html) alongside the SE report. Quarto rendering requires the package to be installed with vignettes built (the default for make install); when the Quarto CLI or the built sources are absent, report rendering is skipped with a warning instead of failing the run.

• DEA report (Grp2Analysis_V2_R6.Rmd and its Quarto port Grp2Analysis_V2_R6_quarto.qmd): the interactive PCA and volcano plots now use a responsive width with a fixed height instead of a fixed pixel width, fixing the tall/narrow rendering; the abundance-density subplot is given an explicit height so it no longer collapses.

• New Quarto vignette Grp2Analysis_V2_R6_quarto.qmd: a Quarto port of the differential-expression report, styled with the vendored FGCZ Quarto template extension and built through the package vignette machinery. It loads the analysis object from a serialized .rds via the deanalyse_file parameter (falling back to a generated example when none is supplied), places the table of contents in a wider right-hand margin, and enables the FGCZ plot-finder toolbar (the floating “Find” and “Save”/download buttons). The existing R Markdown report is unchanged.

• CLI logging: the command-line entry points now route stray message() and warning() output (from prolfqua core, dplyr joins, and prolfquapp itself) through the logger, so those lines appear with the standard INFO/WARN [timestamp] layout and are captured in the run log file instead of printing untagged. readr’s column-specification chatter (Rows:/Columns:, Use spec()) is silenced. New exported helper route_messages_to_logger(). # prolfquapp 2.3.3

• Protein abundance QC report (QC_ProteinAbundances.Rmd): contaminant and decoy proteins are now drawn on top of the regular points instead of being hidden behind them, and are rendered with a separate, higher opacity so they stand out. plot_abundance_vs_percent() gains a highlight_alpha argument (default 1, so existing callers are unchanged); the report sets it to 0.8.

• DiffExpQC report (DiffExpQC_R6.Rmd): reverted the four combined figures from interactive plotly subplots back to static gridExtra::grid.arrange() panels.