Changes in version 0.99.5                        

  - trumpet_plot(), forest_plot() and effect_compare_plot() match column
    names case-insensitively, like as_gwas_data()

                       Changes in version 0.99.4                        

  - Column auto-detection is now case-insensitive, so lowercase headers
    (pos, #chrom, alt, ...) are recognized
  - -log10(p) columns (LOG10P, neg_log_pvalue, ...) are auto-detected
    and back-transformed
  - Fixed a crash when more than one required column could not be
    detected; the error now lists the missing columns

                       Changes in version 0.99.3                        

  - Added GRanges interoperability: as_granges() exports results to a
    Bioconductor GRanges, and as_gwas_data() now accepts GRanges input
  - Added trumpet_plot(): effect size versus minor allele frequency with
    statistical-power contours showing which variants a study can detect
  - Added forest_plot() for effect estimates with confidence intervals
    across cohorts or lead variants
  - Added effect_compare_plot() comparing variant effects between two
    studies on their shared variants
  - Added gene_annotation() with bundled protein-coding genes for GRCh37
    and GRCh38, so regional and gene-labelled plots work without a GTF
  - gene_track() now renders full exon structure when given exon_data
  - Sped up smart_downsample() with exponential-key weighted sampling

                       Changes in version 0.99.2                        

  - Added y_truncate parameter for Manhattan plots with broken y-axis,
    showing extreme p-values in a compressed zone above the break
  - Added snp_density() with heatmap and points styles for SNP density
    karyograms with centromere markers
  - Added density_signal_plot() dual-track comparison of genotyping
    density vs association signal
  - Added chr_info_human(), chr_info_mouse(), chr_info_cattle() for
    built-in chromosome data, and chr_info_ucsc() for any UCSC assembly
  - Used Greek letter on axis labels in volcano and architecture plots
  - Reversed color scale in heatmaps (dark = high values)
  - Filled empty bins in pvalue_heatmap() to eliminate gaps
  - Replaced scattermore with geom_point for correct circle shapes
  - Showed legends by default in journal themes
  - Reduced overlapping text in genetic correlation and PheWAS plots

                       Changes in version 0.99.1                        

  - Added PheWAS, colocalization, fine-mapping, genetic correlation, and
    architecture plots
  - Added data utilities: filter_region(), maf_filter(), merge_gwas(),
    get_loci()
  - Added scattermore support for faster rendering of large datasets
  - Expanded vignette with interpretation guides for all plot types
  - Added Zenodo DOI
  - Fixed integer overflow in cumulative BP calculation
  - Fixed NA handling in highlight_regions()

                       Changes in version 0.99.0                        

Initial Bioconductor pre-release.

Plot types

  - Manhattan plot with smart downsampling, highlighting, and labeling
  - QQ plot with confidence bands, genomic inflation factor, and
    stratification
  - Miami plot for two-study comparison
  - Locus zoom plot with LD coloring and gene track support
  - Genome-wide p-value heatmap (novel)
  - Effect-size volcano plot (novel)
  - Circular Manhattan with multi-ring support (novel)
  - Enrichment Manhattan with functional annotation overlays (novel)
  - Multi-trait Manhattan with pleiotropy detection (novel)
  - Summary dashboard with automatic panel tags (novel)

Gene annotation

  - manhattan_genes() for labeling peaks with gene names
  - annotate_genes() for nearest-gene mapping
  - top_hits() with clumping and cytoband estimation
  - highlight_regions() for marking genomic regions
  - Arrow-style annotations

Themes and palettes

  - Journal themes: Nature, Science, Cell, PLOS, presentation, poster
  - 14 colorblind-safe palettes
  - Publication presets via gwas_preset()

Data I/O

  - Readers for PLINK, REGENIE, GCTA, GEMMA, and generic formats
  - Automatic column name detection
  - as_gwas_data() constructor with validation