LoginSign Up
← Back to all sparks
G

glyvis

ANALYTICS
Velocity0.0

Fast Visualization For Glycoverse

github.com/glycoverse/glyvis ↗

glyvis keeps losing plot functions as the packages behind them get reorganized.

glycomicsvisualizationggplot2bioconductorbreaking changes
◆Current state
glyvis is the plotting layer for glycoverse results. Its recent releases are dominated by two forces it does not control: glyexp's container migration, which it absorbed in 0.7.0 by accepting SummarizedExperiment inputs, and glystats' function removals, which cost it first the WGCNA and consensus-clustering autoplot methods and then the entire enrichment plotting surface. Its own additions in the window are narrow, mostly label handling and NA robustness.
◆Where it's heading
The package is being pruned from upstream rather than expanded from within. Every breaking change in the last four releases is a removal triggered by a sibling package dropping the function that produced the object being plotted. With enrichment now living in glyfun, the plotting for it has to be rebuilt somewhere, and glyvis is the obvious home.
◆Prediction
Expect enrichment plotting to return once glyfun's result objects stabilize, since the visualizations were removed for want of an upstream producer rather than because users stopped needing them.

◆Recent moves

  1. 1mo ago

    Docs recommend the new SE containers

    Documentation-only update pointing users at GlycomicSE and GlycoproteomicSE, closing glyvis's part of the Stage II migration.

    View source ↗
  2. 1mo ago

    plot_logo() detects glycoproteomics across both containers

    plot_logo() recognizes glycoproteomics experiments consistently whether the input is a legacy experiment() or a current container. A small fix, but plot_logo() is the one function with a hard experiment-type requirement, so inconsistent detection broke it outright.

    View source ↗
  3. 1mo ago

    Plots accept SummarizedExperiment; enrichment plots removed

    The visualization methods accept GlycomicSE and GlycoproteomicSE inputs, while the enrichment plotting functions are deleted because glystats removed the gly_enrich_*() functions that fed them. Migration and amputation in the same release.

    View source ↗
  4. 3mo ago

    Stale WGCNA and clustering autoplot methods removed

    The autoplot() methods for glystats WGCNA and consensus-clustering results are removed after the upstream functions disappeared. Dead code cleanup with no effect on working pipelines.

    View source ↗
  5. 6mo ago

    Dependencies move to the r-universe repository

    A packaging change adopting r-universe for dependency resolution, applied across the glycoverse packages in the same week. No user-visible behavior change.

    View source ↗
  6. 7mo ago

    plot_logo() fetches UniProt sequences automatically

    plot_logo() pulls protein sequences from UniProt when no FASTA is supplied, removing a manual preparation step before sequence logo plots. One of the few additions in this window that originated inside glyvis.

    View source ↗