Changes in version 0.99.2 Author identity correction - Corrected the package author and maintainer display name to Shuchao Ren across DESCRIPTION, package help, vignettes, README citation guidance, and Bioconductor submission metadata. - Updated the recommended package citation to: Shuchao Ren (2026). GeneTrackR: Genomic Track Visualization and Haplotype Analysis. R package version 0.99.2. - Retained the GitHub account name Renscq in repository URLs; it is an account identifier rather than the package author name. - Preserved all production R code and analysis behavior from GeneTrackR 0.99.1. Changes in version 0.99.1 Package metadata and citation normalization - Synchronized the package version, title, author identity, release date, README citation guidance, and Bioconductor submission metadata for the 0.99.1 review candidate. - Standardized the recommended package citation as: Shuchao Ren (2026). GeneTrackR: Genomic Track Visualization and Haplotype Analysis. R package version 0.99.1. - Kept package citation generation based on DESCRIPTION metadata instead of adding a custom inst/CITATION, avoiding a BiocCheck warning for a package-only citation without an associated publication DOI. - Preserved all production analysis code, statistical behavior, genomic I/O behavior, tests, vignettes, and generated help content from 0.99.0. Changes in version 0.99.0 Bioconductor submission candidate - Promoted GeneTrackR from the 0.8.x Bioconductor-readiness series to the required 0.99.0 pre-release version for formal new-package submission. - Raised the R dependency floor to R 4.6.0 so the submission candidate matches the current Bioconductor 3.24 devel build environment. - Froze the public analysis APIs and pure-R genomic I/O implementation from GeneTrackR 0.8.9; this release does not introduce new analysis algorithms or change statistical behavior. - Added final submission preparation and preflight tooling for a package-only GitHub default branch, Bioconductor devel validation, and the 2026 BiocContributions review workflow. - Retained NEWS, executable vignettes, runnable help examples, deterministic bundled demo data, GRanges interoperability, and the GPL-3 license declaration required by the submission candidate. Changes in version 0.8.9 Final BiocCheck source cleanup - Added explicit return-value documentation to the advanced-parameter and deterministic-demo help topics so BiocCheck no longer reports missing \value sections. - Moved the GenePred() runnable example into its roxygen source block so devtools::document() preserves the example instead of overwriting a manually synchronized Rd example. - Replaced a stop(paste0(...)) condition with sprintf() and reworded the strict merge-conflict error message to avoid BiocCheck condition-signaling false positives without changing control flow. - Kept network-dependent repository warnings, submission versioning, Support Site registration, ORCID/funding metadata, long-function notices, and package-wide formatting notes outside this source hotfix. Changes in version 0.8.8 Runnable writer example hotfix - Fixed the write_feature_track() runnable example to use bundled GenePred gene-model data converted with as_feature(), matching the wrapper's default BED12 requirement for transcript and exon information. - Added a regression test that exercises the same GenePred-to-Feature-to-BED12 path used by the help example. - Kept write_feature_track() implementation and all annotation-writing behavior unchanged. Changes in version 0.8.7 Documentation and regression hotfix - Linked all as_feature() S3 methods to the shared roxygen topic so devtools::document() no longer creates a nameless as_feature.GenePred topic. - Updated the GenePred coercion example to exercise as_feature() directly with bundled demo annotation data. - Aligned the duplicated-identifier regression test with the 0.8.6 BiocCheck-compliant error path: conflict = "error" now expects a direct error without a preliminary warning. - Kept merge semantics, annotation conversion behavior, and all analysis algorithms unchanged. Changes in version 0.8.6 Bioconductor compliance cleanup - Expanded runnable examples across exported help pages to satisfy the Bioconductor documentation coverage requirement. - Added missing return-value documentation for compatibility writer functions. - Updated the R dependency floor to 4.5.0 and added the suggested VariantAnnotation biocView. - Replaced avoidable 1: sequences, <<- state mutation, and BiocCheck-flagged condition patterns without changing analysis outputs. - Kept large-scale formatting, long-function refactoring, and legitimate warning suppression out of this focused compatibility release. Changes in version 0.8.5 Restore Bioconductor release notes - Restored NEWS.md as a maintained package-level release history because Bioconductor strongly recommends NEWS files and BiocCheck includes dedicated NEWS checks. - Restored the pkgdown News navigation entry and the submission-tree workflow so release notes remain available in both development documentation and the package-only Bioconductor source tree. - Updated the Bioconductor preflight and submission preparation tools to require and preserve NEWS.md. - Kept GeneTrackR analysis code, public APIs, statistical behavior, and genomic I/O unchanged from 0.8.4. Changes in version 0.8.4 R CMD check cleanup - Resolved the remaining package-owned R CMD check warning by documenting all arguments exposed in exported function usage sections. - Declared data.table and ggplot2 non-standard-evaluation symbols reported by R CMD check, eliminating undefined-global-variable diagnostics without changing runtime behavior. - Wrapped long Rd examples so generated manual examples no longer exceed the 100-character width check. - Standardized the package license declaration to GPL-3 and removed the redundant top-level LICENSE file. - The Windows-only future file timestamps NOTE remains an environment time-verification issue rather than a GeneTrackR source problem. Changes in version 0.8.3 Bioconductor submission hygiene - Separate the development workspace from the package-only Bioconductor submission tree. - Add a development-side submission-tree preparation tool and expand the Bioconductor preflight checks for default-branch hygiene. - Label every R code chunk in all package vignettes for current BiocCheck compatibility. - Document bundled inst/extdata files and their deterministic provenance in a package help topic. - Update README installation guidance to lead with the Bioconductor installation path while retaining GitHub development instructions. - Keep GeneTrackR analysis APIs and executable package behavior unchanged. Changes in version 0.8.2 - Hardened package documentation for Bioconductor build/check workflows without changing analysis algorithms or public APIs. - Added author/date metadata, explicit Introduction and Installation sections, package landing-page links, and executable session information to all 15 formal vignettes. - Numbered all 15 VignetteIndexEntry values so browseVignettes() follows the intended GeneTrackR workflow order instead of filename alphabetic order. - Added build-aware vignette rendering: normal R package builds use 144 dpi to reduce CPU, memory, and artifact size, while the pkgdown workflow sets GENETRACKR_PKGDOWN=true to retain 300-dpi website figures. - Kept publication export examples capable of 300-500 dpi output while executing vignette figure exports at the build-aware documentation DPI. - Replaced all \dontrun{} examples in GeneTrackR roxygen/man documentation with compact, deterministic, runnable examples based on bundled gtr_demo_* data and temporary output paths. - Reduced expensive signal examples to the compact RNA-seq demo where frame-specific Ribo-seq behavior was not essential to the function being documented. - Added tools/bioc_preflight.003.R to audit runnable examples, vignette structure, documentation DPI policy, individual file size, estimated source-tree size, and R-build exclusion of development-only directories. Changes in version 0.8.1 - Expanded Bioconductor interoperability so as_granges() now accepts annotation, in-memory signal (BwgTrack), and in-memory variant (VariantTrack) objects while preserving the existing annotation-level interface. - Added retrieve_vcf(..., as = "GRanges") so indexed or in-memory variant retrieval can feed directly into standard Bioconductor genomic interval workflows. - Added S4Vectors, grDevices, and tools to explicit runtime imports where GeneTrackR calls their namespaces directly; retained Rsamtools in Suggests because tabix-backed querying remains an optional acceleration path. - Added dependency-placement and cross-track GRanges regression coverage, including lazy-track conversion diagnostics. - Documented the boundary between GeneTrackR's native pure-R bedGraph/WIG/BigWig I/O and Bioconductor's rtracklayer infrastructure while keeping the native backend and public signal APIs unchanged. - Added tools/bioc_preflight.002.R with dependency duplication, namespace-declaration, and interoperability checks while retaining the 0.8.0 preflight script for development history. Changes in version 0.8.0 - Started the Bioconductor-readiness phase while keeping the 0.7.8 pure-R signal implementation and public analysis APIs unchanged. - Added Bioconductor-oriented DESCRIPTION metadata, including software biocViews, a broader package title, and a three-sentence functional description that reflects annotation, signal, variant, haplotype, phenotype, LD, and variant-effect workflows. - Kept NeedsCompilation: no, removed stale generated DESCRIPTION fields and the obsolete roxygen configuration pin, and added openxlsx to Suggests for the optional workbook workflow used in an executable vignette. - Added RiboParser-style README navigation links for pkgdown documentation, issues, citation, and license, and added explicit Citation and License sections. - Converted all pkgdown article Rmd files into formal R package vignettes with VignetteIndexEntry, VignetteEngine, and UTF-8 metadata so they can participate in R CMD build and R CMD check. - Added a Bioconductor-readiness regression test for metadata and GRanges interoperability and a versioned developer preflight script for repository, vignette, compiled-code, and submission-boundary checks. - Kept the package at version 0.8.0 during preflight development; the formal Bioconductor submission candidate will use the required 0.99.0 version series. Changes in version 0.7.8 - Finalized the 0.7.x pure-R signal migration documentation without changing public APIs or signal calculations. - Removed obsolete Rcpp, bundled libBigWig, src/, and compiled-backend wording from installation guides, signal/export tutorials, pkgdown vignettes, generated README content, and public signal help pages. - Restored README.qmd as the canonical include-based source used by tools/render_readme.R; the clean 0.7.7 deployment artifact had retained the rendered README but stripped its include directives. - Updated package-level documentation to state that bedGraph, WIG, and BigWig reading, regional querying, and writing use the built-in pure-R signal I/O architecture. - Synchronized the public signal Rd pages with their roxygen sources and renamed legacy writer-test labels to match the pure-R implementation. - Retained schema-v2 BwgTrack, 1-based closed internal coordinates, lazy/memory semantics, tabix bedGraph behavior, BigWig chromosome-size requirements, and all 0.7.7 signal behavior unchanged. Changes in version 0.7.7 - Added full signal-subsystem regression coverage for lazy-versus-memory BigWig retrieval, chromosome-edge intervals, strand semantics, normalization parity, and cross-format round trips. - Fixed the pure-R package boundary regression test so data.table key attributes are not mistaken for signal-value differences, while keeping compiled-helper absence as a hard requirement. - Improved the compiled-helper failure diagnostic to identify stale namespace bindings and remind developers to remove R/bw_cpp_backend.R and restart R before testing. - Prevented summary_bwg() and bin_bwg() from mutating caller-owned data.tables through data.table reference semantics. - Fixed merge_bwg(sample_conflict = "keep_first") so duplicate sample IDs retain records and sequence metadata only from the first source, including lazy objects without in-memory signal data. - Retained the pure-R bedGraph, WIG, and BigWig I/O architecture, schema-v2 BwgTrack contract, and public signal APIs unchanged. Changes in version 0.7.6 - Removed the compiled libBigWig/Rcpp compatibility backend after native-R BigWig reading, querying, and writing had been migrated and unified in 0.7.2-0.7.5. - Removed R/bw_cpp_backend.R and the complete src/ tree, including libBigWig C sources, the Rcpp bridge, platform Makevars files, and compiled build artifacts. - Removed Rcpp from Imports and LinkingTo, removed the generated Rcpp import and GeneTrackR useDynLib() registration, and changed package metadata to NeedsCompilation: no. - Replaced the obsolete compiled-reader parity test with native low-level-versus-adapter parity and added package-level regression checks that compiled BigWig helpers are absent from the namespace. - Updated the BigWig writer architecture test to follow the 0.7.5 dispatcher design (write_bwg() -> write_signal_file_memory() -> write_bigwig_native()) instead of requiring a direct writer call from write_bwg(). - Retained the existing public signal APIs, schema-v2 BwgTrack contract, 1-based closed coordinates, tabix bedGraph behavior, and native-R BigWig round-trip behavior. Changes in version 0.7.5 - Added a unified native-R signal I/O layer for bedGraph, WIG, and BigWig formats while preserving the existing public read_bwg(), retrieve_bwg(), and write_bwg() APIs. - Centralized in-memory format dispatch through read_signal_file_memory() and write_signal_file_memory(), and centralized lazy per-sample retrieval through query_signal_file_region(). - Replaced duplicate bedGraph parsing in full-file and tabix query paths with one canonical 0-based half-open to 1-based closed parser. - Reworked WIG text reading around the stabilized bwTools parser pattern, including fixedStep, variableStep, bedGraph-style records, comments, and gzip/bgzip text connections. - Centralized signal-region clipping and sequence-metadata construction so all supported formats follow the same canonical six-column signal schema and schema-v2 seqinfo contract. - Retained compatibility wrappers for existing internal text I/O helper names while removing their duplicate implementations from read_bwg.R and write_bwg.R. - Added regression coverage for unified dispatch, compressed text inputs, interval clipping, and bedGraph/WIG round trips without coordinate drift. - Kept Rcpp and src/ temporarily for compiled-backend parity code only; their final removal remains isolated to the next migration stage. Changes in version 0.7.4 - Fixed schema-v2 seqinfo chromosome subsetting in retrieve_bwg(..., as = "BwgTrack") and slice_bwg(..., as = "BwgTrack"); the previous helper allowed the data.table chrom column to mask the function argument and therefore retained unrelated chromosomes. - Added the native-R BigWig writer adapted from the stabilized bwTools 0.8.10 implementation, including binary packing, chromosome B+ tree construction, R-tree indexing, compressed data blocks, total summaries, and zoom-level generation. - Switched write_bwg(..., format = "bigwig") from the bundled libBigWig/Rcpp writer to the native R writer without changing the public write_bwg() argument contract. - Kept chrom_sizes mandatory for GeneTrackR BigWig export in this release to avoid mixing backend migration with API changes. - Added native writer regression coverage for round-trip signal values, multiple samples, overlap and chromosome-boundary validation, and multi-level chromosome indexes. - Retained the compiled backend only for temporary parity tests and later cleanup; public BigWig reading and writing now both use native R implementations. Changes in version 0.7.3 - Switched the public BigWig read path from the bundled libBigWig/Rcpp reader to the native R reader introduced in 0.7.2. - Added internal query_bigwig_native() and read_bigwig_whole_native() adapters that return the standard GeneTrackR signal schema. - Updated read_bwg() memory loading and sequence metadata collection, lazy retrieve_bwg() queries, and legacy seqinfo_bwg() fallback to use the native R backend. - Updated BwgTrack$meta$backend to GeneTrackR-native-R for newly read signal objects. - Retained the compiled reader temporarily only for regression comparison; BigWig writing remains on the compiled backend until the writer migration. - Added regression coverage that verifies public BigWig read functions no longer reference the compiled reader while preserving interval values, clipping, and schema behavior. Changes in version 0.7.2 - Added the first native-R BigWig reader layer adapted from the stabilized bwTools 0.8.10 implementation. - Added internal binary decoding for BigWig headers, chromosome B+ trees, R-tree indexes, compressed data blocks, and 32-bit floating-point signal values. - Added internal bigwig_metadata_native(), bigwig_seqinfo_native(), and bigwig_query_native() functions using 1-based closed coordinates without Rcpp or .Call(). - Added metadata caching keyed by normalized path, file size, and modification time while keeping the native reader local-file only. - Added a deterministic BigWig fixture and regression tests for chromosome metadata, interval decoding, clipping, empty ranges, invalid files, and parity with the existing compiled backend. - Kept read_bwg() and retrieve_bwg() on the existing libBigWig/Rcpp backend in this release so the new reader can be verified before the public backend switch in 0.7.3. Changes in version 0.7.1 - Established the schema-v2 BwgTrack object foundation for the 0.7.x pure-R signal subsystem migration. - Added a stable seqinfo slot alongside samples, data, meta, and validation; chromosome lengths can be known or NA when the source format does not encode sequence lengths. - Standardized BwgTrack metadata to coordinate = "1-based closed" and schema_version = "2" while preserving the existing public constructor argument order. - Added sequence metadata collection for bigWig inputs and in-memory bedGraph/WIG inputs, and preserved relevant seqinfo through BwgTrack retrieval, slicing, and merging. - Updated seqinfo_bwg() to use stored schema-v2 sequence metadata first while retaining the existing libBigWig fallback for legacy objects. - Added regression tests for schema construction, sequence metadata validation, text-track sequence metadata, subset propagation, and merge propagation. - Kept the existing libBigWig/Rcpp backend unchanged in this release; native BigWig replacement begins in the following 0.7.x development stages. Changes in version 0.6.17 - Optimized pkgdown article figure layouts by increasing the graphics canvas according to plot complexity while keeping rendered figures at full article width. - Standardized vignette web graphics to 150 dpi with fig.retina = 1; publication-oriented export examples continue to use 500 dpi. - Reduced documentation-only plotting text sizes to 9-10 pt for functions that expose text_size or font, without changing GeneTrackR plotting defaults or package APIs. - Enlarged annotation-specific figure heights and retained all R implementations, staged docs/*.qmd sources, tests, and statistical behavior unchanged. Changes in version 0.6.16 - Removed the Documentation QA GitHub Actions workflow. - Retained the pkgdown workflow as the single documentation build and deployment workflow. - Kept README rendering, pkgdown articles, package code, tests, and generated documentation unchanged. Changes in version 0.6.15 - Fixed pkgdown homepage code rendering by generating README.md from README.qmd and staged docs/*.qmd sources as display-only GitHub-Flavored Markdown; Quarto/knitr \``{r}fences are converted to standard```r` fences instead of being exposed as plain text. - Added tools/render_readme.R as the single deterministic README assembly entry point. The script expands Quarto include directives without executing the 202 documentation code chunks and supports --check for CI synchronization checks. - Added README synchronization to Documentation QA and made the pkgdown workflow regenerate README.md before site construction, preventing stale homepage content. - No R implementation, API, vignette, statistical method, or plotting behavior was changed. Changes in version 0.6.14 - Removed the noisy GitHub Pages 404 annotation from the pkgdown workflow by probing repository Pages availability before invoking actions/configure-pages. - actions/configure-pages now runs only when the Pages REST endpoint is available; repositories without Pages enabled receive a clean pkgdown preview artifact and Step Summary without an Error or Warning annotation. - Kept pkgdown site construction as a hard CI check and retained automatic native Pages deployment once repository-level Pages is enabled. - No R implementation, API, vignette, test, or pkgdown content was changed. Changes in version 0.6.13 - Synchronized the package-level Rd file with the roxygen2 8.1.0 output generated from DESCRIPTION URL and BugReports metadata. - Added the automatically generated Useful links \seealso{} section so the documentation QA roxygen drift check remains clean. - No R implementation, API, vignette, test, pkgdown workflow, or analysis behavior was changed. Changes in version 0.6.12 - Made the pkgdown GitHub Pages deployment tolerant of repositories where Pages has not yet been enabled: pkgdown site construction remains a hard check, while Pages configuration/deployment is skipped with a clear warning instead of failing the entire workflow. - Added a fallback pkgdown-site-preview workflow artifact when GitHub Pages is unavailable, so the generated site can still be inspected from the Actions run. - Retained the native GitHub Pages artifact/deploy path automatically when actions/configure-pages succeeds; no R implementation, API, vignette, or pkgdown content was changed. Changes in version 0.6.11 - Re-synchronized roxygen2 8.1.0 generated metadata, NAMESPACE, and package-level Rd documentation after the 0.6.10 GitHub commit omitted part of the generated-file update. - Retained the documentation QA roxygen drift guard and the explicit roxygen2@8.1.0 CI pin. Changes in version 0.6.10 - Synchronized roxygen-generated package metadata, NAMESPACE, and package-level Rd documentation with roxygen2 8.1.0, matching the generator used by the documentation QA workflow. - Pinned roxygen2 8.1.0 in the documentation-QA dependency set so future CI runs do not fail solely because a newer roxygen2 release reformats generated documentation. - Preserved all R implementations, tests, pkgdown articles, staged Quarto sources, and package APIs unchanged; this release is limited to generated-documentation synchronization and CI reproducibility. Changes in version 0.6.9 - Added a dedicated GitHub Actions documentation-QA workflow that rebuilds the pkgdown site in a clean R process, verifies roxygen-generated documentation is synchronized, checks internal site links, inventories generated figure dimensions, and runs R CMD check --as-cran. - Added a Bioconductor release preflight for main-branch and manually triggered runs using the official bioconductor/bioconductor_docker:RELEASE_3_23 image and BiocCheck; package-specific documentation/coding errors fail CI, while submission-only biocViews checks are deferred until the formal Bioconductor submission phase. - Modernized pkgdown deployment to the native GitHub Pages artifact workflow while retaining pkgdown-site/ as the generated-site directory. - Kept the 0.6.8 article, Reference, API, and analysis implementations unchanged; this release is limited to documentation QA and CI integration. Changes in version 0.6.8 - Added a dedicated export and reproducibility article covering standard genomic track writers, PDF/PNG figure export, tabular result preservation, LD matrix export, optional openxlsx workbooks, complete RDS objects, manifests, and sessionInfo() provenance. - Added a production-oriented complete workflow article connecting annotation, interval features, RNA-seq/Ribo-seq signals, variants, haplotypes, phenotype association, LD, phenotype-guided refinement, variant-effect prioritization, and reproducible export through one consistent candidate-locus analysis. - Added a Workflow & reproducibility section to the pkgdown Articles index while retaining core-workflows as the shorter Getting Started workflow. - Kept the staged docs/13-export.qmd and docs/14-workflow.qmd sources unchanged while migrating their current workflows into formal pkgdown vignettes. Changes in version 0.6.7 - Added a dedicated linkage-disequilibrium article covering genotype-dosage LD calculation, LDTrack result inspection, r2 and dosage-based D-prime interpretation, triangular heatmaps, regional gene/variant context, population subsets, variant filtering, and downstream regional haplotype handoff. - Added a dedicated phenotype-guided haplotype-refinement article covering original-versus-refined haplotype state, pairwise merge evidence, effect_threshold, multi-trait grouping, grouped and collapsed refined-variant plots, and refined phenotype visualization. - Added a dedicated variant-effect article covering binary genotype coding, absolute versus signed effects, adjusted P values, linked tied-effect interpretation, direct single-variant phenotype validation, and multi-trait prioritization. - Extended the pkgdown Haplotype & association article sequence to Haplotype -> Phenotype -> LD -> Haplotype refinement -> Variant effect while keeping the staged docs/10-ld.qmd, docs/11-haplotype-refinement.qmd, and docs/12-variant-effect.qmd sources unchanged. Changes in version 0.6.6 - Added a dedicated haplotype article covering gene-, transcript-, and region-defined HapVariant construction, genotype representations, missing-genotype filtering, sample/variant filtering, object inspection, and plot_hap_variant() visualization. - Added a dedicated phenotype-association article covering phenotype import and QC, explicit sample-ID alignment checks, haplotype-phenotype testing, multi-trait inference, single-variant phenotype testing, and preservation of returned statistical tables. - Added a task-oriented Haplotype & association section to the pkgdown Articles index, positioned after variant/browser visualization and before the later LD/refinement/variant-effect modules. - Kept the staged docs/08-haplotype.qmd and docs/09-phenotype.qmd sources unchanged while migrating their current workflows into pkgdown vignettes. Changes in version 0.6.5 - Added a dedicated pkgdown variant-track article covering VariantTrack construction, memory versus indexed lazy VCF access, validation, summaries, genomic/gene/transcript retrieval, ID/type/pattern filtering, variant plotting, site-level VCF export, and downstream handoff. - Added a dedicated integrated browser-track article covering plot_tracks() locator semantics, gene/transcript/region views, strand-aware RNA-seq and Ribo-seq display, transcript-level Ribo-seq frame rendering, feature and variant panels, panel layout, highlighting, styling, and large-dataset usage. - Extended the pkgdown Visualization section from gene models and signal tracks to the complete gene-model -> signal -> variant -> integrated-browser documentation sequence. - Kept all staged docs/*.qmd sources unchanged for subsequent haplotype, phenotype, LD, refinement, variant-effect, export, and end-to-end workflow migration. Changes in version 0.6.4 - Added a dedicated pkgdown gene-model visualization article covering gene, transcript, and regional structural views, genomic versus spliced transcript coordinates, transcript collapse, strand-direction arrows, labels, highlights, and explicit feature colors. - Rebuilt the signal-track article around the complete BwgTrack workflow: reading, lazy versus in-memory access, strand semantics, regional retrieval, normalization, merging, binning, writing, gene/transcript/region visualization, Ribo-seq frame views, and integrated plot_tracks() figures. - Separated gene-model structure from continuous-signal documentation so later variant/browser-track articles can build on a stable visualization foundation without duplicating annotation or signal basics. - Added the gene-model article to the pkgdown Visualization section while retaining all staged docs/*.qmd sources unchanged for subsequent migration steps. Changes in version 0.6.3 - Added a dedicated pkgdown annotation article covering GenePred/GenePredExt, GTF, GFF3, and BED inputs through one unified GeneTrackR workflow. - Reorganized annotation documentation around native coordinate handling, object contracts, validation and summaries, feature retrieval, cross-format conversion, merging, writing, and round-trip checks. - Added annotation-focused visualization examples for genes, transcripts, genomic regions, interval feature tracks, and gene-length distributions. - Added a Data & annotation section to the pkgdown Articles index while retaining the existing Getting Started and Visualization navigation. - Kept all staged docs/*.qmd sources unchanged for later module-specific migration. Changes in version 0.6.2 - Rebuilt the pkgdown Get Started path around a dedicated first-run article that covers installation, package loading, bundled demo inputs, and a minimal annotation-to-haplotype/phenotype workflow. - Added a deterministic example-data article documenting the shared file inventory, coordinate conventions, core dimensions, sample alignment rule, GeneA haplotype/phenotype truth, LD gradient, and strand-specific RNA-seq/Ribo-seq design. - Updated the pkgdown article order to Getting started -> Example data -> Core workflow, while retaining signal-track documentation under Visualization. - Added prerequisite navigation to the existing core workflow without changing its executable analysis workflow. - Kept all staged docs/*.qmd sources unchanged for later module-by-module migration. Changes in version 0.6.1 - Redesigned the pkgdown information architecture around a dedicated Get Started entry, task-oriented Articles, and a functionally grouped Reference index. - Organized all current Rd topics into explicit Reference sections for core objects, annotation I/O and conversion, gene-model visualization, signal tracks, variant tracks, haplotype analysis, phenotype association, linkage disequilibrium, variant effect/refinement, and advanced diagnostics. - Positioned core-workflows as the top-level Get Started workflow and retained signal-tracks under the Visualization article section. - Added article descriptions to the existing vignettes so the pkgdown article index explains the purpose of each workflow. - Kept the staged docs/*.qmd tutorial sources unchanged; additional task-oriented articles will be migrated into the established information architecture in subsequent 0.6.x releases. Changes in version 0.6.0 - Initialized the GeneTrackR package website with pkgdown, replacing the abandoned MkDocs migration plan with an R-native documentation architecture. - Added _pkgdown.yml with the canonical GitHub Pages URL, Bootstrap 5, the existing workflow vignettes, and a temporary pkgdown-site/ output directory. - Added a GitHub Actions workflow to build the package website on pushes, pull requests, releases, and manual dispatches, and to deploy non-PR builds to the gh-pages branch. - Added package website and issue-tracker metadata to DESCRIPTION, together with the website-only pkgdown dependency declaration. - Preserved the existing docs/*.qmd documentation sources during staged migration; they remain the source material used by README.qmd and are not overwritten by pkgdown in this release. Changes in version 0.5.37 - Fixed haplotype allele fill color helpers so the stable A, T, C, G, and indel mapping retains its names after alpha adjustment. - Fixed variant marker color helpers so the stable SNP, Ind, and ... mapping retains its names after alpha adjustment. - Updated plotting-default regression tests to preserve the intentional Reds default for plot_ld_block() while other plotting palettes remain Paired. - Changed full-memory bigWig loading to emit one memory warning per read_bwg() call instead of one warning per input file, and made round-trip tests explicitly expect this warning. Changes in version 0.5.36 - Rebuilt docs/13-export.qmd as a structured export workflow for genomic tracks, PDF/PNG figures, analysis tables, optional Excel workbooks, and complete RDS objects. - Rebuilt docs/14-workflow.qmd as the end-to-end GeneTrackR candidate-locus workflow using the current GeneA haplotype, phenotype, LD-gradient, refinement, and variant-effect examples. - Removed docs/15-notes.qmd and incorporated its large-data and object-return guidance into the end-to-end workflow. - Synchronized README.qmd and README.md with the revised documentation sequence. Changes in version 0.5.35 - Fixed plot_variant_effect() multi-trait reshaping so mixed integer and double phenotype columns no longer trigger data.table::melt() coercion warnings. - Standardized selected numeric traits to double before long-format conversion without changing their values or effect statistics. - Added regression coverage requiring the documented multi-trait variant-effect workflow to run silently. - Clarified mixed numeric trait handling in docs/12-variant-effect.qmd. - Bumped the package version to 0.5.35. Changes in version 0.5.34 - Rebuilt docs/12-variant-effect.qmd as a complete regional variant-effect prioritization workflow. - Added explicit guidance for binary genotype coding, absolute versus signed effects, adjusted P values, and multi-trait interpretation. - Documented the deterministic GeneA effect truth: the p13-linked variant set has an effect of approximately +6 on protein_content, varA05 has an intermediate effect, and varA02 is a zero-effect control. - Clarified that tied phenotype effects among linked variants do not identify a unique causal variant and should be interpreted together with LD, annotation, direct variant-phenotype plots, and refinement. - Added regression coverage for the redesigned variant-effect documentation workflow. - Bumped the package version to 0.5.34. Changes in version 0.5.33 - Replaced docs/10-ld.qmd with the user-updated LD workflow while restoring valid executable QMD syntax. - Redesigned the GeneA genotype/phenotype truth so refinement follows genotype-nearest haplotype pairs. - Changed varA03 from the cross-cluster p23 pattern to the cluster-consistent p13 pattern. - Redesigned phenotype means so Hap1/Hap2 are locally similar, Hap3/Hap4 are locally similar, and cross-cluster differences are larger. - Updated the protein_content refinement truth to Hap1 + Hap2 versus Hap3 + Hap4. - Updated phenotype, refinement, variant-effect, workflow, vignette, generator, validator, help examples, and regression tests. - Bumped the package version to 0.5.33. Changes in version 0.5.32 - Expanded the primary LD demo from 6 to 12 natural SNPs while preserving the 11-variant GeneA haplotype truth. - Redesigned the primary LD demo to contain a perfect high-LD core plus progressively decorrelated downstream variants, producing a clear r2 gradient with both high- and low-LD pairs. - Changed the default plot_ld_block() heatmap palette from Paired to Reds. - Updated LD documentation, demo truth metadata, generated VCF data, and regression tests. Changes in version 0.5.31 - Rebuilt docs/10-ld.qmd as a complete LD calculation, inspection, visualization, and downstream handoff workflow. - Added deterministic high-LD, low-LD, and two-variant demo examples with explicit expected pairwise behavior. - Documented the LDTrack object contract, pairwise statistics, LD matrix, genotype retention, sample/variant filtering, and plot_ld_block() return semantics. - Clarified that GeneTrackR calculates pairwise LD but does not automatically call biological block boundaries from a fixed threshold. - Added guidance on dosage-based Dprime limitations for unphased VCF data and recommended r2 as the primary generic metric. - Added documentation regression coverage for the designed LD truth sets. Changes in version 0.5.30 - Rebuilt docs/11-haplotype-refinement.qmd as a phenotype-guided refinement workflow. - Changed the primary demo refinement trait to protein_content, which deterministically refines four GeneA haplotypes into two phenotype groups. - Added explicit inspection of pairwise tests, merge decisions, trait-specific groups, original-to-refined mappings, grouped/collapsed refined variant plots, and refined phenotype plots. - Added guidance on effect_threshold, trait dependence, negative-control refinement, and multi-trait refinement. - Added documentation regression coverage for the designed 4 -> 2, 4 -> 4, and 4 -> 1 refinement outcomes. Changes in version 0.5.29 - Reorganized docs/09-phenotype.qmd into a self-contained phenotype-association workflow covering phenotype QC, sample matching, haplotype association, multiple testing, single-variant association, result objects, and downstream reuse. - Documented the designed demo phenotype roles for seed_weight, plant_height, protein_content, flowering_time, and flower_color. - Clarified numeric versus categorical phenotype support and the relationship among raw p-values, adjusted p-values, and displayed significance brackets. - Added documentation regression checks for sample-ID matching and the phenotype workflow. Changes in version 0.5.28 - Reorganized docs/08-haplotype.qmd into a self-contained, stepwise haplotype workflow from VCF input through downstream analysis handoff. - Documented the HapVariant object structure, genotype representations, missing-genotype filtering, flanking-region semantics, sample/variant filtering, and haplotype plotting controls. - Established the GeneA gene body as the primary tutorial haplotype definition (11 variants, 36 samples, four balanced haplotypes) and separated the upstream missing-genotype example from the main analysis. Changes in version 0.5.27 - Added variant_palette and variant_colors controls to plot_tracks() and forwarded them to integrated variant panels. - Standardized all plotting palette defaults to Paired, including signal, frame, feature, variant, LD, haplotype-table, and phenotype fill palettes. - Updated browser-track documentation and regression tests for variant color control and package-wide palette defaults. Changes in version 0.5.26 - Improved browser-track feature legends by adding automatic compact feature grouping in plot_tracks() and plot_feature_track(). - Added transcript-centered automatic Ribo-seq frame rendering in plot_tracks() via ribo_signal_type = "auto". - Updated docs/07-browser-tracks.qmd to explain feature legend grouping and transcript-centered Ribo-seq frame behavior. Changes in version 0.5.25 - Reorganized docs/07-browser-tracks.qmd into a self-contained browser-track workflow covering input preparation, gene/transcript/region locators, signal selection, feature/variant integration, layout, heights, highlighting, and large-data usage. - Clarified the exact-one-locator rule used by plot_tracks() and the distinction between browser-level genomic signal plotting and transcript-coordinate/frame-specific signal visualization. - Updated browser examples to use the current public read_vcf() API and explicit strand/sample selection. - Documented current panel ordering, named feature/variant list behavior, and the compact styling scope of integrated feature/variant tracks. Changes in version 0.5.24 - Fix fixed-string pattern filtering so fixed = TRUE no longer passes the incompatible ignore.case argument to grepl(). - Preserve case-insensitive fixed-string matching by normalizing both the pattern and candidate fields before matching. - Add regression tests for warning-free fixed matching and case-insensitive fixed-string VCF retrieval. Changes in version 0.5.23 - Fix summary_vcf() so an in-memory VariantTrack can be summarized without supplying a genomic region. - Fix retrieve_vcf() so region, gene, and transcript filters are optional for in-memory/full-file queries; ID, type, and pattern filters can now be used independently as documented. - Add validation for incomplete direct regions and flank arguments used without a gene/transcript locator. - Add regression tests for whole-object summaries and non-regional VCF retrieval. Changes in version 0.5.22 - Reorganized docs/06-variant.qmd into a continuous VCF workflow covering reading, validation, summaries, region/gene/transcript retrieval, ID/type/pattern filtering, and plotting. - Clarified VariantTrack versus data.table return semantics for retrieve_vcf(). - Added explicit guidance for indexed lazy VCF access and strand-aware gene/transcript retrieval. - Documented the current site-level scope of write_vcf() to avoid implying genotype-preserving round-trip export. Changes in version 0.5.21 - Changed discrete signal palette assignment so sample, group, and frame levels use RColorBrewer class colors in strict level order instead of interpolating between palette endpoints; this also fixes plot_signal_region() color ordering. - Kept heatmap palettes continuous while separating continuous-gradient generation from discrete signal color assignment. - Reworked demo Ribo-seq counts so frame 0, frame 1, and frame 2 have visibly heterogeneous heights and different occupied codons while preserving approximately 80%/10%/10% total counts. - Increased frame-0 codon occupancy while reducing off-frame occupancy, producing a dense but visually varied RPF pattern; initiation and termination frame-0 counts remain approximately two-fold over the internal mean. - Added regression checks for discrete palette order and within-frame Ribo-seq count variability. Changes in version 0.5.20 - Fixed signal qualitative palettes so discrete levels use RColorBrewer colors strictly in palette order; frame0, frame1, and frame2 now map to the first three Set1 colors by default. - Reworked demo Ribo-seq counts to moderate density with approximately 80%/10%/10% frame0/frame1/frame2 total-count proportions. - Reduced demo Ribo-seq initiation and termination peaks to approximately two times the internal frame-0 mean.